A long, medium and short term nested cascade hydropower station forecast scheduling system

Abstract

The application provides a long-medium-short period nested cascade hydropower station forecast scheduling system, which comprises a long-medium-short period nested power generation optimization scheduling module and a reservoir inflow quantity forecast module; the long-medium-short period nested power generation optimization scheduling module comprises a long-term optimization scheduling module of the cascade hydropower station, a medium-term optimization scheduling module of the cascade hydropower station, a short-term optimization scheduling module of the cascade hydropower station, an intra-plant economic operation module of the cascade hydropower station and a scheduling model heuristic optimization algorithm module; the reservoir inflow quantity forecast module comprises a monthly inflow quantity prediction module based on a self-made probability distribution model, a daily inflow quantity LSTM prediction module based on multidimensional feature data driving and an hourly inflow quantity prediction module. The economic operation module is embedded, the cascade hydropower station operation is optimized, the efficient management of time and space scales is realized, the new algorithm is adopted to correct the initial value of the flow, the water quantity balance and the tail water level constraint are met, the optimization efficiency is improved, the data driving prediction is utilized to forecast the inflow quantity, the model calculation is simplified and the forecast scheduling efficiency is improved.

Description

Technical Field

[0001] This invention relates to the field of hydropower technology, specifically to a forecasting and scheduling system for cascade hydropower stations with nested long, medium, and short-term forecasts. Background Technology

[0002] The optimal operation and management of cascade small hydropower groups are becoming increasingly complex due to the influence of random factors such as watershed meteorology and hydrology. Considering the impact of uncertain factors such as inflow and load, scholars have made numerous achievements in the field of reservoir stochastic optimal scheduling and forecasting scheduling, starting from two aspects: utilization of scheduling information and construction of optimal scheduling models. Common methods include three categories: explicit stochastic optimal scheduling, implicit stochastic optimal scheduling, and hierarchical nested optimal scheduling.

[0003] In practical engineering applications, optimized scheduling must consider both runoff variations during the current period and short-term benefits, as well as the long-term evolution of runoff to improve long-term benefits. Furthermore, in addition to random runoff, uncertainties such as load fluctuations also significantly impact scheduling decisions. Therefore, the actual operation of a reservoir is a rolling optimization decision-making process of "forecasting, decision-making, implementation, re-forecasting, re-decision-making, and re-implementation." Existing explicit and implicit random optimization scheduling methods are insufficient to meet the actual needs of reservoir optimized operation and management. Hierarchical nested optimization scheduling technology, however, can provide rolling optimization decisions and better guide actual reservoir operation. Nevertheless, existing hierarchical nested optimization scheduling systems need to be organically integrated with the plant's economic operation module. Summary of the Invention

[0004] The main objective of this invention is to provide a nested forecasting and scheduling system for cascade hydropower stations with long, medium, and short-term timeframes, addressing the aforementioned problems.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] A nested long-term, medium-term, and short-term cascade hydropower station forecasting and scheduling system includes a nested long-term, medium-term, and short-term power generation optimization scheduling module and a reservoir inflow forecasting module;

[0007] The nested long-term, medium-term, and short-term power generation optimization scheduling module includes: a long-term optimization scheduling module for cascade hydropower stations, a medium-term optimization scheduling module for cascade hydropower stations, a short-term optimization scheduling module for cascade hydropower stations, an in-plant economic operation module for cascade hydropower stations, and a scheduling model heuristic optimization algorithm module.

[0008] The reservoir inflow forecasting module includes: a monthly inflow forecasting module based on a self-made probability distribution model, a daily inflow LSTM forecasting module driven by multidimensional feature data, and an hourly inflow forecasting module.

[0009] While adopting the above technical solutions, the present invention may also adopt or combine the following technical solutions:

[0010] As a preferred embodiment of the present invention, the operation of the long-term, medium-term, and short-term nested power generation optimization scheduling module includes the following steps:

[0011] S1: Initialize the target end water level in December of year k and use it as the end water level boundary constraint of the long-term optimization scheduling module of the cascade hydropower station, k = current year, l = 1;

[0012] S2: Execute the long-term optimal scheduling module for the cascade hydropower stations and call the scheduling model heuristic optimization algorithm module for optimization, to obtain the corrected optimal long-term scheduling retention period reference flow and the end-of-month water level of all months within the long-term scheduling retention period, i.e., the end-of-month water level of the reservoir from month 1 of year k to month 12 of year k. The objective function of the long-term optimal scheduling module for the cascade hydropower stations is:

[0013]

[0014] In the formula, T y For the long-term scheduling carryover period, i.e., from the current month to the 12th month, Δt y The duration of the long-term scheduling period is 1 month, n s and n re These represent the total number of power stations and the total number of reservoirs in a cascade hydropower project, respectively. i t , Let be the total power output and the flow rate of the i-th power station in time period t, respectively. For the kth re The discharge flow of each reservoir in time period t;

[0015] S3: The water level at the end of the first month of the long-term scheduling remainder period, i.e. the water level at the end of the reservoir in the lth month of the kth year, is used as the boundary constraint of the final water level of the mid-term optimization scheduling module of the cascade hydropower station.

[0016] S4: Initialize m = 1;

[0017] S5: Execute the mid-term optimization scheduling module for the cascade hydropower stations and call the scheduling model heuristic optimization algorithm module for optimization, to obtain the corrected optimal mid-term scheduling remainder period reference flow and the end-of-day water level of all days within the mid-term scheduling remainder period, i.e., the end-of-day water level of the reservoir from the mth day of the 1st month of the kth year to the end of the 1st month of the kth year. The objective function of the mid-term optimization scheduling module for the cascade hydropower stations is:

[0018]

[0019] In the formula, T m The medium-term scheduling carryover period is from the current day to the end of the month, Δt m The duration of the intermediate scheduling period is 1 day;

[0020] S6: The water level at the end of the first day of the medium-term optimization scheduling period, i.e. the water level at the end of the day of the reservoir on the mth day of the lth month of the kth year, is used as the boundary constraint of the end water level of the short-term optimization scheduling module of the cascade hydropower station.

[0021] S7: Execute the short-term optimal scheduling module for the cascade hydropower stations and call the scheduling model heuristic optimization algorithm module for optimization to obtain the corrected optimal daily drawdown flow, daily drawdown flow plan, daily power generation plan, and the hourly water level for all hours within the day, i.e., the hourly water level of the reservoir from the 1st hour of the 1st day of the 1st month of the kth year to the 24th hour of the 1st day of the 1st month of the kth year. The objective function of the short-term optimal scheduling module for the cascade hydropower stations is:

[0022]

[0023] In the formula, T d For short-term scheduling, the scheduling period is 1 day, Δt d For short-term scheduling periods, i.e., 1 hour, c t The electricity price for generation in time period t;

[0024] S8: Initialize n = 1;

[0025] S9: Determine whether to adopt the "electricity-determined water supply" method. If yes, execute the cascade hydropower station's internal economic operation module and call the scheduling model heuristic optimization algorithm module for optimization, carrying out the "electricity-determined water supply" cascade hydropower station's internal economic operation to obtain the optimal unit output. If no, execute the cascade hydropower station's internal economic operation module and call the scheduling model heuristic optimization algorithm module for optimization, carrying out the "water-determined electricity supply" cascade hydropower station's internal economic operation to obtain the optimal unit flow rate. If the "electricity-determined water supply" method is adopted, the objective function of the cascade hydropower station's internal economic operation module is:

[0026]

[0027] If the "water-based power generation" approach is adopted, the objective function of the economic operation module within the cascade hydropower station is:

[0028]

[0029] In the formula, q ij P ij Let n be the unit's reference flow rate and unit output of the j-th generating unit of the i-th power station, respectively, and Δt be the duration of the current time period, i.e., 1 hour. u This represents the total number of generating units in the power plant.

[0030] S10: Determine if n≥24? If yes, proceed to the next step; otherwise, go to S14.

[0031] S11: Determine if m ≥ the total number of days in month l. If yes, proceed to the next step; otherwise, go to S15.

[0032] S12: Determine if l≥12? If yes, proceed to the next step; otherwise, go to S16.

[0033] S13: Determine if the process is complete. If yes, end the process; otherwise, proceed to S17.

[0034] S14: Determine if the time is a new hour. If yes, then n = n + 1 and go to S9. If no, continue to determine the time.

[0035] S15: Determine if the time is a new day. If yes, then m = m + 1 and go to S5. If no, continue to determine the time.

[0036] S16: Determine if the time is a new month. If yes, then l = l + 1 and go to S2. If no, continue to determine the time.

[0037] S17: Determine if the time is the beginning of a new year. If yes, go to S1; otherwise, continue determining the time.

[0038] As a preferred technical solution of the present invention: the daily reference flow plan = the corrected optimal daily reference flow, and the daily power output plan = the power plant comprehensive power output coefficient * the corrected optimal daily reference flow * the head.

[0039] As a preferred embodiment of the present invention: the long-term scheduling remainder period reference flow population correction value, the medium-term scheduling remainder period reference flow population correction value, and the daily reference flow population correction value, as well as the corrected optimal long-term scheduling remainder period reference flow, the corrected optimal medium-term scheduling remainder period reference flow, and the corrected optimal daily reference flow, are all obtained through a forward reservoir water level calculation method containing a reverse reference flow correction algorithm and include the following steps:

[0040] S101: Initialize the upper limit of reservoir capacity V max Lower limit of reservoir capacity V min Maximum traffic limit Q e_max The maximum reference traffic Q for the first time period e1_max Total number of time periods NT, maximum reference traffic Q e1_max The expression is:

[0041] Q e1_max =[SL L (1)-(V min -V b (1))] / DT(1)-Q s (1) (6)

[0042] In the formula, SL L (1) V b (1), DT(1) and Q s(1) The reservoir inflow, initial reservoir capacity, duration of the period, and ecological flow are respectively for the first period.

[0043] S102: Determine the reference traffic Q in the first time period. e (1) > Q e1_max If yes, then Q e (1) = Q e1_max If yes, proceed to the next step; otherwise, proceed directly to the next step.

[0044] S103: Initialize t k =1;

[0045] S104: Determine t k <NT-1? If yes, proceed to the next step; otherwise, execute the backreference traffic correction algorithm and go to S108.

[0046] S105: Determine V b (t k )+SL L (t k )-(Q e (t k )+Q s (t k ))*DT(t k )>V max If yes, then

[0047]

[0048] If not, proceed to step S107;

[0049] In the formula, V b (t k V e (t k V x (t k ),SL L (t k ), Q e (t k ), Q s (t k ) and DT(t k ) are respectively the tth k The reservoir's initial capacity, final capacity, discharge, inflow, diversion flow, ecological flow, and duration for each time period;

[0050] S106: Determine V b (t k )+SL L (t k )-(Q e (t k)+Q s (t k ))*DT(t k ) < V min If yes, then V e (t k ) = V min If not, execute the backreference traffic correction algorithm and proceed to the next step.

[0051]

[0052] And proceed to the next step;

[0053] S107: The tth k +1 hour initial water level of reservoir V b (t k +1)=V e (t k ), t k =t k +1, and switch to S104;

[0054] S108: The reservoir's final storage capacity V for all time periods is calculated using the water level-storage capacity relationship curve. e Convert the water level to the final water level of the reservoir for all time periods, and then end the process.

[0055] As a preferred embodiment of the present invention, the reverse reference traffic correction algorithm includes the following steps:

[0056] S111: Initialize t k2 =t k ;

[0057] S112: Determine t k2 ≥1? If yes, proceed to the next step; otherwise, go to S117.

[0058] S113: Determine V e (t k2 )-V b (t k2 ) <SL L (t k2 )-Q s (t k2 )*DT(t k2 If yes, then...

[0059]

[0060] If not, proceed to step S117;

[0061] In the formula, V b (t k2 V e (t k2V x (t k2 ),SL L (t k2 ), Q e (t k2 ), Q s (t k2 ) and DT(t k2 ) are respectively the tth k2 The reservoir's initial capacity, final capacity, discharge, inflow, diversion flow, ecological flow, and duration for each time period;

[0062] S114: Determine t k2 =1? Yes, then calculate the initial water level V of the reservoir in the first time period according to the following formula. b1

[0063]

[0064] Proceed to the next step; otherwise, proceed to S116.

[0065] S115: Judgment | V b (t k2 )-V b1 | / V b (t k2 If ) < 1E-5? Yes, then

[0066]

[0067] If not, an error occurs, the reference flow cannot be corrected, and the process is redirected to S117.

[0068]

[0069] t k2 =t k2 -1, and switch to S112;

[0070] In the formula, V e (t k2 -1) represents the t-th... k2 -1 water level at the end of the reservoir period;

[0071] S117: Backreference traffic correction terminated, process ended.

[0072] As a preferred embodiment of the present invention, the operation of the monthly water volume prediction module based on the self-made probability distribution model includes the following steps:

[0073] S201: Determine if the time is the beginning of a new year. If yes, proceed to the next step; otherwise, continue to determine the time.

[0074] S202: Statistics on monthly rainfall from January to December of previous years;

[0075] S203: Calculate the probability distribution model of rainfall from January to December respectively. The probability distribution model of rainfall includes the probability density function and probability distribution function of monthly rainfall.

[0076] S204: Calculate the mathematical expectation of the probability distribution model of rainfall from January to December, and use it as the monthly rainfall forecast for the corresponding month of the current year;

[0077] S205: Calculate the 90% confidence intervals of the probability distribution models of rainfall from January to December, and use them as the monthly rainfall prediction intervals for the corresponding months of the current year.

[0078] S206: Calculate the predicted monthly inflow and the predicted range of monthly inflow for the current year based on the relationship curve between monthly inflow and monthly rainfall.

[0079] S207: Determine if the process is complete. If yes, end the process; otherwise, proceed to S201.

[0080] As a preferred technical solution of the present invention: the daily water volume LSTM prediction module driven by multi-dimensional feature data is executed once at the beginning of each month.

[0081] As a preferred technical solution of the present invention: the hourly water inflow prediction module is executed once a day.

[0082] This invention provides a nested forecasting and scheduling system for cascade hydropower stations with long, medium, and short-term implications, offering the following advantages: It embeds an in-plant economic operation module within the forecasting and scheduling process, achieving efficient optimization of cascade hydropower station operation at monthly to hourly time scales and at spatial scales from the station level to the unit level; in the process of optimizing the reference flow during forecasting and scheduling, a forward reservoir water level calculation method with a reverse reference flow correction algorithm corrects the initial value of the reference flow randomly generated by the heuristic algorithm, ensuring that the corrected reference flow simultaneously satisfies water balance constraints and final water level boundary constraints, thereby effectively reducing the probability of solutions that do not meet the constraints and improving optimization efficiency; it uses a data-driven approach to predict the inflow volume for the next year, month, and day, ensuring forecast accuracy while eliminating the need to construct complex hydrological models and perform inflow process evolution calculations, thus improving the implementation efficiency of forecasting and scheduling. Attached Figure Description

[0083] Figure 1 This is a schematic diagram of the nested long-term, medium-term, and short-term cascade hydropower station forecasting and scheduling system provided by the present invention.

[0084] Figure 2 The flowchart shows the operation of the nested long-term, medium-term, and short-term power generation optimization scheduling module.

[0085] Figure 3 This is a flowchart of a forward reservoir water level calculation method that includes a reverse reference flow correction algorithm.

[0086] Figure 4 This is a flowchart of the reverse reference traffic correction algorithm.

[0087] Figure 5 This is a flowchart of the monthly water volume prediction module based on a self-made probability distribution model.

[0088] Figure 6 A schematic diagram of the hydraulic connections of the Xiecunyuan cascade hydropower station.

[0089] Figure 7 This is a graph showing the relationship between monthly inflow and monthly rainfall at Xiecunyuan Second-Level Reservoir.

[0090] Figure 8 This is a chart showing the predicted monthly rainfall for the Xiecunyuan Second-Level Reservoir.

[0091] Figure 9 This is a chart showing the predicted monthly water inflow to Xiecunyuan Second-Level Reservoir.

[0092] Figure 10 This is a chart showing the predicted daily inflow to Xiecunyuan Second-Level Reservoir.

[0093] Figure 11 The output curves of Unit 1 of Xiecunyuan Second-Level Hydropower Station before and after optimization.

[0094] Figure 12 The output curves of Unit 2 of Xiecunyuan Second-Level Hydropower Station before and after optimization.

[0095] Figure 13 The output curves of Unit 1 of Xiecunyuan Third-Level Hydropower Station before and after optimization.

[0096] Figure 14 The output curves of Unit 2 of Xiecunyuan Third-Level Hydropower Station before and after optimization.

[0097] In the diagram: 1-Long-term, medium-term, and short-term nested power generation optimization scheduling module; 2-Reservoir inflow forecasting module; 11-Long-term optimization scheduling module for cascade hydropower stations; 12-Medium-term optimization scheduling module for cascade hydropower stations; 13-Short-term optimization scheduling module for cascade hydropower stations; 14-Economic operation module within the cascade hydropower plant; 15-Heuristic optimization algorithm module for scheduling models; 21-Monthly inflow prediction module based on a self-made probability distribution model; 22-Daily inflow LSTM prediction module driven by multi-dimensional feature data; 23-Hourly inflow prediction module. Detailed Implementation

[0098] The present invention will be described in further detail with reference to the accompanying drawings and specific embodiments.

[0099] like Figure 1 As shown, a nested long-term, medium-term, and short-term cascade hydropower station forecasting and scheduling system includes a nested long-term, medium-term, and short-term power generation optimization scheduling module 1 and a reservoir inflow forecasting module 2.

[0100] The nested long-term, medium-term, and short-term power generation optimization scheduling module 1 includes: a long-term optimization scheduling module 11 for cascade hydropower stations, a medium-term optimization scheduling module 12 for cascade hydropower stations, a short-term optimization scheduling module 12 for cascade hydropower stations, an in-plant economic operation module 14 for cascade hydropower stations, and a scheduling model heuristic optimization algorithm module 15.

[0101] The reservoir inflow forecast module 2 includes: a monthly inflow forecast module 21 based on a self-made probability distribution model, a daily inflow LSTM forecast module 22 based on multidimensional feature data, and an hourly inflow forecast module 23.

[0102] The data interaction relationships between modules are as follows:

[0103] The long-term optimal scheduling module for cascade hydropower stations obtains the predicted monthly inflow value from the monthly inflow prediction module based on a self-developed probability distribution model as input. It calculates the corrected value of the long-term residual flow population based on the original value of the population during the long-term scheduling residual period, and then calculates the objective function value for the long-term scheduling residual period, inputting it into the scheduling model heuristic optimization algorithm module. The scheduling model heuristic optimization algorithm module iteratively updates the original value of the long-term residual flow population and sends it back to the long-term optimal scheduling module for the cascade hydropower stations for a new round of calculations. This process is repeated until the maximum number of iterations is reached, and the optimal solution for the final generation is then sent back to the long-term optimal scheduling module for the cascade hydropower stations.

[0104] The mid-term optimization scheduling module of the cascade hydropower station obtains the predicted daily inflow value as input from the daily inflow LSTM prediction module driven by multi-dimensional feature data, and obtains the water level at the end of the first month of the long-term scheduling remainder as the boundary constraint from the long-term optimization scheduling module. It calculates the corrected value of the reference flow population for the mid-term scheduling remainder based on the original value of the population, and then calculates the objective function value for the mid-term scheduling remainder and inputs it into the scheduling model heuristic optimization algorithm module. The scheduling model heuristic optimization algorithm module iteratively updates the new original value of the reference flow population for the mid-term scheduling remainder and sends it back to the mid-term optimization scheduling module of the cascade hydropower station for a new round of calculation. This process is repeated until the maximum number of iterations is reached, and the optimal solution of the final generation is sent to the mid-term optimization scheduling module of the cascade hydropower station.

[0105] The short-term optimal scheduling module for cascade hydropower stations obtains hourly inflow forecasts from the hourly inflow forecast module as input, and uses the last water level of the first day of the medium-term scheduling remainder as the boundary constraint from the medium-term optimal scheduling module. It calculates the daily reference flow population correction value based on the original daily reference flow population value, and then calculates the daily objective function value, which is input into the scheduling model heuristic optimization algorithm module. The scheduling model heuristic optimization algorithm module iteratively updates the new original daily reference flow population value and sends it back to the short-term optimal scheduling module for a new round of calculation. This process is repeated until the maximum number of iterations is reached, and the final optimal solution is sent to the short-term optimal scheduling module.

[0106] The cascade hydropower station's internal economic operation module obtains the daily power output plan and daily water consumption plan from the cascade hydropower station's short-term optimal scheduling module as input conditions. It calculates the internal economic operation objective function value based on the unit output population or unit water consumption population and inputs it into the scheduling model heuristic optimization algorithm module. The scheduling model heuristic optimization algorithm module iteratively updates the unit output population or unit water consumption population and sends it back to the cascade hydropower station's internal economic operation module for a new round of calculations. This process is repeated until the maximum number of iterations is reached, and the final optimal solution is sent to the cascade hydropower station's internal economic operation module.

[0107] The hourly water volume prediction module obtains the daily water volume prediction value from the daily water volume LSTM prediction module driven by multi-dimensional feature data, and uses it to calculate the hourly water volume prediction value.

[0108] like Figure 2 As shown, the operation of the long-term, medium-term, and short-term nested power generation optimization scheduling module includes the following steps:

[0109] S1: Initialize the target end water level in December of year k and use it as the end water level boundary constraint of the long-term optimization scheduling module of the cascade hydropower station, k = current year, l = 1;

[0110] S2: Execute the long-term optimal scheduling module for cascade hydropower stations and call the scheduling model heuristic optimization algorithm module for optimization, to obtain the corrected optimal long-term scheduling retention period reference flow and the end-of-month water level of all months within the long-term scheduling retention period, i.e., the end-of-month water level of the reservoir from month 1 of year k to month 12 of year k. The objective function of the long-term optimal scheduling module for cascade hydropower stations is:

[0111]

[0112] In the formula, T y For the long-term scheduling carryover period, i.e., from the current month to the 12th month, Δt y The duration of the long-term scheduling period is 1 month, n s and n reThese represent the total number of power stations and the total number of reservoirs in a cascade hydropower project, respectively. i t , Let be the total power output and the flow rate of the i-th power station in time period t, respectively. For the kth re The discharge flow of each reservoir in time period t;

[0113] S3: The water level at the end of the first month of the long-term scheduling remainder period, i.e. the water level at the end of the reservoir in the lth month of the kth year, is used as the boundary constraint of the final water level of the mid-term optimization scheduling module of the cascade hydropower station.

[0114] S4: Initialize m = 1;

[0115] S5: Execute the mid-term optimization scheduling module for the cascade hydropower stations and call the scheduling model heuristic optimization algorithm module for optimization, to obtain the corrected optimal mid-term scheduling remainder period reference flow and the end-of-day water level for all days within the mid-term scheduling remainder period, i.e., the end-of-day water level of the reservoir from the m-th day of the l-th month of year k to the end of the l-th month of year k. The objective function of the mid-term optimization scheduling module for the cascade hydropower stations is:

[0116]

[0117] In the formula, T m The medium-term scheduling carryover period is from the current day to the end of the month, Δt m The duration of the intermediate scheduling period is 1 day;

[0118] S6: The water level at the end of the first day of the medium-term optimization scheduling period, i.e. the water level at the end of the day of the reservoir on the mth day of the lth month of the kth year, is used as the boundary constraint of the end water level of the short-term optimization scheduling module of the cascade hydropower station.

[0119] S7: Execute the short-term optimal scheduling module for cascade hydropower stations and call the scheduling model heuristic optimization algorithm module for optimization to obtain the corrected optimal daily drawdown flow, daily drawdown flow plan, daily power generation plan, and all hourly water levels within the day, i.e., the hourly water levels of the reservoir from the 1st hour of the 1st day of the 1st month of the kth year to the 24th hour of the 1st day of the 1st month of the kth year. The objective function of the short-term optimal scheduling module for cascade hydropower stations is:

[0120]

[0121] In the formula, T d For short-term scheduling, the scheduling period is 1 day, Δt d For short-term scheduling periods, i.e., 1 hour, c t The electricity price for generation in time period t;

[0122] S8: Initialize n = 1;

[0123] S9: Determine whether to adopt the "electricity-determined water supply" method. If yes, execute the cascade hydropower station's internal economic operation module and call the scheduling model heuristic optimization algorithm module for optimization, carrying out the "electricity-determined water supply" cascade hydropower station's internal economic operation to obtain the optimal unit output. If no, execute the cascade hydropower station's internal economic operation module and call the scheduling model heuristic optimization algorithm module for optimization, carrying out the "water-determined electricity supply" cascade hydropower station's internal economic operation to obtain the optimal unit flow rate. If the "electricity-determined water supply" method is adopted, the objective function of the cascade hydropower station's internal economic operation module is:

[0124]

[0125] If the "water-based power generation" approach is adopted, the objective function of the economic operation module within the cascade hydropower station is:

[0126]

[0127] In the formula, q ij P ij Let n be the unit's reference flow rate and unit output of the j-th generating unit of the i-th power station, respectively, and Δt be the duration of the current time period, i.e., 1 hour. u This represents the total number of generating units in the power plant.

[0128] S10: Determine if n≥24? If yes, proceed to the next step; otherwise, go to S14.

[0129] S11: Determine if m ≥ the total number of days in month l. If yes, proceed to the next step; otherwise, go to S15.

[0130] S12: Determine if l≥12? If yes, proceed to the next step; otherwise, go to S16.

[0131] S13: Determine if the process is complete. If yes, end the process; otherwise, proceed to S17.

[0132] S14: Determine if the time is a new hour. If yes, then n = n + 1 and go to S9. If no, continue to determine the time.

[0133] S15: Determine if the time is a new day. If yes, then m = m + 1 and go to S5. If no, continue to determine the time.

[0134] S16: Determine if the time is a new month. If yes, then l = l + 1 and go to S2. If no, continue to determine the time.

[0135] S17: Determine if the time is the beginning of a new year. If yes, go to S1; otherwise, continue determining the time.

[0136] Daily Flow Rate Plan = Corrected Optimal Daily Flow Rate; Daily Power Output Plan = Power Plant Comprehensive Output Coefficient * Corrected Optimal Daily Flow Rate * Head.

[0137] The long-term optimization scheduling module for cascade hydropower stations operates on a one-year scheduling cycle with a one-month scheduling period. It optimizes the power station's flow rate to maximize total power generation and minimize total water wastage, executing once at the beginning of each month. The module includes a long-term scheduling residual period objective function, constraints, input conditions, and decision variables. Constraints include equality and inequality constraints. Equality constraints include: water balance constraints, water level-reservoir capacity curves, water level-flow curves, initial water level boundary constraints, and final water level boundary constraints. Inequality constraints include: upper and lower limits for water level, upper and lower limits for water head, upper and lower limits for power station flow rate, and upper and lower limits for total power station output. The initial water level boundary constraint is the measured initial water level for the current month, and the final water level boundary constraint is the target final water level for the 12th month. The input condition is the predicted monthly inflow. The decision variable is the long-term scheduling residual period flow rate. Long-term scheduling retention period reference traffic It is a column vector consisting of the monthly drawdown flows of all cascade hydropower stations during the long-term scheduling retention period, expressed as follows:

[0138]

[0139] The mid-term optimization scheduling module for cascade hydropower stations operates on a one-month scheduling cycle with a one-day scheduling period. It optimizes the flow rate used by the power stations, aiming to maximize total power generation and minimize total water wastage, and executes once daily. The module includes the objective function for the mid-term scheduling residual period, constraints, input conditions, and decision variables.

[0140] The difference between the constraints and those of the long-term optimization scheduling module for cascade hydropower stations is that the initial water level boundary constraint of the medium-term optimization scheduling module for cascade hydropower stations is the actual measured initial water level on the day, and the final water level boundary constraint is the water level at the end of the first month of the long-term scheduling retention period.

[0141] The input condition is the predicted daily water volume.

[0142] The decision variable is the reference flow during the medium-term scheduling retention period. Mid-term scheduling retention period reference traffic It is a column vector consisting of the daily drawdown flows of all cascade hydropower stations during the medium-term scheduling retention period, expressed as follows:

[0143]

[0144] The short-term optimal scheduling module for cascade hydropower stations operates on a daily scheduling cycle with a 1-hour scheduling period. It optimizes the flow rate used by the power stations, aiming to maximize total power generation revenue and minimize total water wastage. It is executed once daily. The module includes a short-term scheduling objective function, constraints, input conditions, and decision variables. The constraints differ from those in the long-term optimal scheduling module in that the initial water level boundary constraint in the short-term module is the measured initial water level of the day, while the final water level boundary constraint is the final water level of the first day of the medium-term scheduling retention period. The input condition is the hourly inflow forecast.

[0145] The decision variable is daily citation traffic. Daily traffic It is a column vector consisting of the hourly flow rates of all power stations in the cascade hydropower station series within a single day, expressed as follows:

[0146]

[0147] The cascade hydropower station's internal economic operation module is optimized using either a "power-driven water consumption" or "water-driven power consumption" approach. It aims to minimize the total water consumption or maximize the total power generation of the cascade hydropower station during the current time period, and is executed once per hour. The cascade hydropower station's internal economic operation module includes the objective function, constraints, input conditions, and decision variables.

[0148] The constraints include equality constraints and inequality constraints. Equality constraints include: water balance constraints, water level-storage capacity relationship curves, water level-flow relationship curves, and unit efficiency models. Inequality constraints include: upper and lower limits of water level, upper and lower limits of water head, upper and lower limits of unit reference flow, upper and lower limits of unit output, and unit vibration zone constraints.

[0149] The input conditions are the measured inflow, daily power generation plan, and daily water consumption plan for the current period. For the "electricity-determined water supply" cascade hydropower station's economic operation, the unit output allocation is optimized based on the total planned power output for the current period in the daily power generation plan, ensuring that the total power output equals the planned total power output for the current period, thus minimizing the total water consumption of the cascade hydropower station in the current period. For the "water-determined power supply" cascade hydropower station's economic operation, the unit water consumption allocation is optimized based on the planned water consumption for the current period in the daily water consumption plan, ensuring that the total power generation of the cascade hydropower station in the current period is maximized when the power consumption equals the planned water consumption for the current period.

[0150] If the "electricity-based water supply" method is adopted, the decision variable is the unit output; if the "water-based electricity supply" method is adopted, the decision variable is the unit's flow rate.

[0151] Because the process by which the heuristic optimization algorithm module of the scheduling model generates the original values ​​of the long-term, medium-term, and daily reference flow populations is random, when the long-term, medium-term, and short-term optimization scheduling limits the final water level boundary constraints, the aforementioned original values ​​cannot accurately deduce the final water level boundary constraints while satisfying the water balance constraints. This means that the original values ​​of the long-term, medium-term, and daily reference flow populations given by the heuristic optimization algorithm module of the scheduling model cannot simultaneously satisfy both the water balance constraints and the final water level boundary constraints. Therefore, it is necessary to perform reference flow correction based on the final water level boundary constraints so that the corrected reference flow can simultaneously satisfy both the water balance constraints and the final water level boundary constraints. Furthermore, during the long-term, medium-term, and short-term optimization scheduling process, the final generation's population optimal solution output by the scheduling model's heuristic optimization algorithm module is also obtained by optimizing the original values ​​of the reference flow population in each generation. Therefore, it is also necessary to obtain the corrected optimal long-term scheduling residual reference flow, the corrected optimal medium-term scheduling residual reference flow, and the corrected optimal daily reference flow through correction.

[0152] like Figure 3 As shown, the population-corrected values ​​of the long-term scheduling remainder reference flow, the medium-term scheduling remainder reference flow, and the daily reference flow, as well as the corrected optimal long-term scheduling remainder reference flow, the corrected optimal medium-term scheduling remainder reference flow, and the corrected optimal daily reference flow, are all obtained through a forward reservoir water level calculation method containing a reverse reference flow correction algorithm, and include the following steps:

[0153] S101: Initialize the upper limit of reservoir capacity V max Lower limit of reservoir capacity V min Maximum traffic limit Q e_max The maximum reference traffic Q for the first time period e1_max Total number of time periods NT, maximum reference traffic Q e1_max The expression is:

[0154] Q e1_max =[SL L (1)-(V min -V b (1))] / DT(1)-Q s (1) (6)

[0155] In the formula, SL L (1) V b (1), DT(1) and Q s (1) The reservoir inflow, initial reservoir capacity, duration of the period, and ecological flow are respectively for the first period.

[0156] S102: Determine the reference traffic Q in the first time period. e (1) > Q e1_max If yes, then Q e (1) = Q e1_max If yes, proceed to the next step; otherwise, proceed directly to the next step.

[0157] S103: Initialize t k =1;

[0158] S104: Determine t k <NT-1? If yes, proceed to the next step; otherwise, execute the backreference traffic correction algorithm and go to S108.

[0159] S105: Determine V b (t k )+SL L (t k )-(Q e (t k )+Q s (t k ))*DT(t k )>V max If yes, then

[0160]

[0161] If not, proceed to step S107;

[0162] In the formula, V b (t k V e (t k V x (t k ),SL L (t k ), Q e (t k ), Q s (t k ) and DT(t k ) are respectively the tth k The reservoir's initial capacity, final capacity, discharge, inflow, diversion flow, ecological flow, and duration for each time period;

[0163] S106: Determine V b (t k )+SL L (t k )-(Q e (t k )+Q s (t k ))*DT(t k ) < Vmin If yes, then V e (t k ) = V min If not, execute the backreference traffic correction algorithm and proceed to the next step.

[0164]

[0165] And proceed to the next step;

[0166] S107: The tth k +1 hour initial water level of reservoir V b (t k +1)=V e (t k ), t k =t k +1, and switch to S104;

[0167] S108: The reservoir's final storage capacity V for all time periods is calculated using the water level-storage capacity relationship curve. e Convert the water level to the final water level of the reservoir for all time periods, and then end the process.

[0168] like Figure 4 As shown, the backreference traffic correction algorithm includes the following steps:

[0169] S111: Initialize t k2 =t k ;

[0170] S112: Determine t k2 ≥1? If yes, proceed to the next step; otherwise, go to S117.

[0171] S113: Determine V e (t k2 )-V b (t k2 ) <SL L (t k2 )-Q s (t k2 )*DT(t k2 If yes, then...

[0172]

[0173] If not, proceed to step S117;

[0174] In the formula, V b (t k2 V e (t k2 V x (t k2 ),SL L (tk2 ), Q e (t k2 ), Q s (t k2 ) and DT(t k2 ) are respectively the tth k2 The reservoir's initial capacity, final capacity, discharge, inflow, diversion flow, ecological flow, and duration for each time period;

[0175] S114: Determine t k2 =1? Yes, then calculate the initial water level V of the reservoir in the first time period according to the following formula. b1

[0176] V b1 =V e (t k2 )-(SL L (t k2 )-Q s (t k2 )*DT(t k2 (10)

[0177] Proceed to the next step; otherwise, proceed to S116.

[0178] S115: Judgment | V b (t k2 )-V b1 | / V b (t k2 If ) < 1E-5? Yes, then

[0179]

[0180] If not, an error occurs, the reference flow cannot be corrected, and the process is redirected to S117.

[0181] S116:

[0182] t k2 =t k2 -1, and switch to S112;

[0183] In the formula, V e (t k2 -1) represents the t-th... k2 -1 water level at the end of the reservoir period;

[0184] S117: Backreference traffic correction terminated, process ended.

[0185] like Figure 5 As shown, the operation of the monthly water inflow prediction module based on the self-made probability distribution model includes the following steps:

[0186] S201: Determine if the time is the beginning of a new year. If yes, proceed to the next step; otherwise, continue to determine the time.

[0187] S202: Statistics on monthly rainfall from January to December of previous years;

[0188] S203: Calculate the probability distribution model of rainfall from January to December respectively. The probability distribution model of rainfall includes the probability density function and probability distribution function of monthly rainfall.

[0189] S204: Calculate the mathematical expectation of the probability distribution model of rainfall from January to December, and use it as the monthly rainfall forecast for the corresponding month of the current year;

[0190] S205: Calculate the 90% confidence intervals of the probability distribution models of rainfall from January to December, and use them as the monthly rainfall prediction intervals for the corresponding months of the current year.

[0191] S206: Calculate the predicted monthly inflow and the predicted range of monthly inflow for the current year based on the relationship curve between monthly inflow and monthly rainfall.

[0192] S207: Determine if the process is complete. If yes, end the process; otherwise, proceed to S201.

[0193] A daily water inflow prediction module driven by multidimensional feature data is executed once at the beginning of each month to predict the daily water inflow from the 1st to the last day of the month. The prediction is performed using a pre-trained LSTM neural network model. During training, the sample features are {daily maximum temperature, daily minimum temperature, daily rainfall, and the corresponding solar term}, and the result label is daily water inflow. The time expansion step is 30 days, and the forecast period is 1 day. That is, the LSTM neural network model uses the {daily maximum temperature, daily minimum temperature, daily rainfall, and the corresponding solar term} from the 29th day prior to the forecast date to the forecast date itself to fit the daily water inflow for the forecast date. When using the pre-trained LSTM neural network model to predict daily water inflow, the {daily maximum temperature, daily minimum temperature, daily rainfall, and the corresponding solar term} for future days are obtained from weather forecasts.

[0194] The hourly water inflow prediction module is executed once a day to predict the hourly water inflow from the first hour to the last hour of the day, which is obtained by averaging the daily water inflow prediction value over 24 hours.

[0195] Specifically, the aforementioned nested long-, medium-, and short-term cascade hydropower station forecasting and dispatching system is implemented in the following manner:

[0196] Simulation calculations were conducted based on the Xiecunyuan cascade hydropower station. A schematic diagram of the hydraulic connections of the Xiecunyuan cascade hydropower station is shown below. Figure 6As shown, the Xiecunyuan Second-Level Hydropower Station directly draws water from the Xiecunyuan Second-Level Reservoir for power generation. The tailwater of the Xiecunyuan Second-Level Hydropower Station, along with the discharge and ecological flow from the Xiecunyuan Second-Level Reservoir, is then diverted through a weir to the Xiecunyuan Third-Level Hydropower Station as its reference flow. The installed capacities of the Xiecunyuan Second-Level and Xiecunyuan Third-Level Hydropower Stations are 2*8.67MW and 2*2MW, respectively, with rated heads of 219.24m and 48.289m. The electricity price is 0.57 yuan per kilowatt-hour (peak hours: 7:00-11:00 and 13:00-23:00) and 0.228 yuan per kilowatt-hour (valley hours: 11:00-13:00 and 23:00-07:00 the next day).

[0197] The main inequality constraints for the Xiecunyuan cascade hydropower station are shown in Table 1.

[0198] Table 1. Parameter settings for inequality constraints at the Xiecunyuan cascade hydropower station.

[0199]

[0200] The economic operation module within the cascade hydropower station adopts the "water-determined power generation" approach. The scheduling model heuristic optimization algorithm module is based on the adaptive fuzzy particle swarm algorithm (AFPSO algorithm) for iterative optimization. For details of this algorithm, please refer to the literature "Accurate Parameter Estimation of a Hydro-Turbine Regulation System Using Adaptive Fuzzy Particle Swarm Optimization" (Authors: Dong Liu et al., Journal: Energies, Volume: 2019, Issue: 12, Article No: 3903).

[0201] The monthly inflow-monthly rainfall relationship curve of Xiecunyuan Second-Level Reservoir is as follows: Figure 7 As shown, the two exhibit a strong linear relationship, i.e., monthly inflow = β0 + β1 * monthly rainfall + ε, where β0 and β1 are the 0th and 1st order coefficients, respectively, and ε is the residual. Through analysis of... Figure 7 By performing least squares estimation on the scattered points, we can obtain β0 = -5.457E5 and β1 = 5.138E4, with a coefficient of determination of 0.8952.

[0202] Monthly rainfall and monthly inflow at Xiecunyuan Second-Level Reservoir are as follows: Figure 8-9 As shown in the figure, the predicted curve trend is consistent with the actual curve trend, and the 90% confidence interval can basically cover all actual values. At this time, the root mean square error of the monthly water volume prediction is 3.35E6 m. 3 .

[0203] The daily water volume of Xiecunyuan Second-Level Reservoir is as follows Figure 10 As shown in the figure, the predicted curve and the actual curve trend are consistent, and the deviation between the predicted and actual values ​​is small except for some peaks. At this time, the root mean square error of the daily water inflow prediction is 2.25E5 m. 3 .

[0204] Table 2 shows the forecasting and scheduling optimization results of the Xiecunyuan cascade hydropower station based on the embodiments of the present invention. As can be seen from the table, after optimization by the embodiments of the present invention, the annual power generation revenue of the Xiecunyuan cascade hydropower station in 2020 can be increased by 7.8%.

[0205] Table 2 Optimization results of the embodiments of the present invention for Xiecunyuan cascade hydropower station

[0206]

[0207] Drawing as Figure 11-14 The output curves of Unit 1, Unit 2, Unit 1, Unit 2, Unit 1, and Unit 2 of the Xiecunyuan Second-Level Hydropower Station before and after optimization, as well as the output curves of Unit 1 and Unit 2 of the Xiecunyuan Third-Level Hydropower Station before and after optimization, from March 1, 2020 to March 31, 2020, are shown to intuitively demonstrate the plant-wide economic operation effect of the embodiment of the present invention. As shown in the figure, this embodiment of the invention utilizes the in-plant economic operation module 14 of the cascade hydropower station to rationally allocate the load, ensuring maximum total output of the Xiecunyuan cascade hydropower station while avoiding the units operating in the vibration zone. When the total output of the power station is greater than the maximum output of one unit but significantly less than the sum of the maximum outputs of the two units (see the curves from March 13 to March 28), the optimized unit output is not evenly distributed, but one unit operates at full capacity while the other unit carries a partial load, to avoid the two units operating in the vibration zone simultaneously due to an even distribution of the load. When the total output of the power station is close to the sum of the maximum outputs of the two units (see the curves after March 29), after optimization, the two units distribute the load evenly, ensuring that neither operates in the vibration zone. Furthermore, when the total output of the power station is significantly less than the maximum output of one unit (see the curves from March 6 to March 8), after optimization, one unit operates under partial load conditions in the non-vibration zone, while the other unit is shut down or operates at low load.

[0208] The above specific embodiments are used to explain and illustrate the present invention, and are only preferred embodiments of the present invention, not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made to the present invention within the spirit and scope of the claims shall fall within the protection scope of the present invention.

Claims

1. A nested long-term, medium-term, and short-term forecasting and dispatching system for cascade hydropower stations, characterized in that: This includes a nested short-term, medium-term, and long-term power generation optimization scheduling module and a reservoir inflow forecasting module; The nested long-term, medium-term, and short-term power generation optimization scheduling module includes: a long-term optimization scheduling module for cascade hydropower stations, a medium-term optimization scheduling module for cascade hydropower stations, a short-term optimization scheduling module for cascade hydropower stations, an in-plant economic operation module for cascade hydropower stations, and a scheduling model heuristic optimization algorithm module. The reservoir inflow forecasting module includes: a monthly inflow forecasting module based on a self-made probability distribution model, a daily inflow LSTM forecasting module driven by multi-dimensional feature data, and an hourly inflow forecasting module. The operation of the nested long-term, medium-term, and short-term power generation optimization scheduling module includes the following steps: S1: Initialize the target end water level in December of year k and use it as the end water level boundary constraint of the long-term optimization scheduling module of the cascade hydropower station, k = current year, l = 1; S2: Execute the long-term optimal scheduling module for the cascade hydropower stations and call the scheduling model heuristic optimization algorithm module for optimization, to obtain the corrected optimal long-term scheduling retention period reference flow and the end-of-month water level of all months within the long-term scheduling retention period, i.e., the end-of-month water level of the reservoir from month 1 of year k to month 12 of year k. The objective function of the long-term optimal scheduling module for the cascade hydropower stations is: （1） In the formula, This is a long-term scheduling reserve period, from the current month to the 12th month. The duration of the long-term scheduling period is one month. and These represent the total number of power stations and the total number of reservoirs in a cascade hydropower project. , Let be the total power output and the flow rate of the i-th power station in time period t, respectively. For the first The discharge flow of each reservoir in time period t; S3: The water level at the end of the first month of the long-term scheduling remainder period, i.e. the water level at the end of the reservoir in the lth month of the kth year, is used as the boundary constraint of the final water level of the mid-term optimization scheduling module of the cascade hydropower station. S4: Initialize m=1; S5: Execute the mid-term optimization scheduling module for the cascade hydropower stations and call the scheduling model heuristic optimization algorithm module for optimization, to obtain the corrected optimal mid-term scheduling remainder period reference flow and the end-of-day water level of all days within the mid-term scheduling remainder period, i.e., the end-of-day water level of the reservoir from the mth day of the 1st month of the kth year to the end of the 1st month of the kth year. The objective function of the mid-term optimization scheduling module for the cascade hydropower stations is: （2） In the formula, This is the medium-term scheduling reserve period, which is from the current day to the end of the month. The duration of the intermediate scheduling period is 1 day; S6: The water level at the end of the first day of the medium-term optimization scheduling period, i.e. the water level at the end of the day of the reservoir on the mth day of the lth month of the kth year, is used as the boundary constraint of the end water level of the short-term optimization scheduling module of the cascade hydropower station. S7: Execute the short-term optimal scheduling module for the cascade hydropower stations and call the scheduling model heuristic optimization algorithm module for optimization to obtain the corrected optimal daily drawdown flow, daily drawdown flow plan, daily power generation plan, and the hourly water level for all hours within the day, i.e., the hourly water level of the reservoir from the 1st hour of the 1st day of the 1st month of the kth year to the 24th hour of the 1st day of the 1st month of the kth year. The objective function of the short-term optimal scheduling module for the cascade hydropower stations is: （3） In the formula, The scheduling cycle is short-term, i.e., 1 day. The duration of the short-term scheduling period is 1 hour. The electricity price for generation in time period t; S8: Initialize n=1; S9: Determine whether the "electricity-determined water supply" method is adopted. If yes, execute the cascade hydropower station's internal economic operation module and call the scheduling model heuristic optimization algorithm module to find the optimal unit output by conducting the "electricity-determined water supply" cascade hydropower station's internal economic operation. If no, execute the cascade hydropower station's internal economic operation module and call the scheduling model heuristic optimization algorithm module to find the optimal unit flow rate by conducting the "water-determined electricity supply" cascade hydropower station's internal economic operation. If the "electricity-determined water supply" method is adopted, the objective function of the cascade hydropower station's internal economic operation module is: （4） If the "water-based power generation" approach is adopted, the objective function of the economic operation module within the cascade hydropower station is: （5） In the formula, , These represent the unit's reference flow rate and unit output, respectively, for the j-th generating unit of the i-th power station. The duration of the current time period is 1 hour. This represents the total number of generating units in the power plant. S10: Determine if n≥24? If yes, proceed to the next step; otherwise, go to S14. S11: Determine if m ≥ the total number of days in month l. If yes, proceed to the next step; otherwise, go to S15. S12: Determine if l≥12? If yes, proceed to the next step; otherwise, go to S16. S13: Determine if the process is complete. If yes, end the process; otherwise, proceed to S17. S14: Determine if the time is a new hour. If yes, then n = n + 1 and go to S9. If no, continue to determine the time. S15: Determine if the time is a new day. If yes, then m = m + 1 and go to S5. If no, continue to determine the time. S16: Determine if the time is a new month. If yes, then l = l + 1 and go to S2. If no, continue to determine the time. S17: Determine if the time is the new year. If yes, go to S1; otherwise, continue to determine the time. The operation of the monthly water volume prediction module based on the self-made probability distribution model includes the following steps: S201: Determine if the time is the beginning of a new year. If yes, proceed to the next step; otherwise, continue to determine the time. S202: Statistics on monthly rainfall from January to December of previous years; S203: Calculate the probability distribution model of rainfall from January to December respectively. The probability distribution model of rainfall includes the probability density function and probability distribution function of monthly rainfall. S204: Calculate the mathematical expectation of the probability distribution model of rainfall from January to December, and use it as the monthly rainfall forecast for the corresponding month of the current year; S205: Calculate the 90% confidence intervals of the probability distribution models of rainfall from January to December, and use them as the monthly rainfall prediction intervals for the corresponding months of the current year. S206: Calculate the predicted monthly inflow and the predicted range of monthly inflow for the current year based on the relationship curve between monthly inflow and monthly rainfall. S207: Determine if the process is complete. If yes, end the process; otherwise, proceed to S201.

2. The nested long-term, medium-term, and short-term cascade hydropower station forecasting and dispatching system according to claim 1, characterized in that: The daily reference flow plan = the corrected optimal daily reference flow, and the daily power output plan = the power plant's comprehensive power output coefficient * the corrected optimal daily reference flow * the head.

3. The nested long-term, medium-term, and short-term cascade hydropower station forecasting and dispatching system according to claim 1, characterized in that: The long-term remaining period reference flow population correction value, the medium-term remaining period reference flow population correction value, and the daily reference flow population correction value, as well as the corrected optimal long-term remaining period reference flow, the corrected optimal medium-term remaining period reference flow, and the corrected optimal daily reference flow, are all obtained through a forward reservoir water level calculation method containing a reverse reference flow correction algorithm and include the following steps: S101: Initialize the upper limit of reservoir capacity Lower limit of reservoir capacity Maximum number of references The maximum number of references allowed in the first time period. Total number of time periods Maximum number of references The expression is: （6） In the formula, , , and These are the reservoir inflow, initial reservoir capacity, duration of the period, and ecological flow for the first time period. S102: Determine the reference traffic in the first time period. If yes, then If yes, proceed to the next step; otherwise, proceed directly to the next step. S103: Initialization ; S104: Judgment If yes, proceed to the next step; otherwise, execute the backreference traffic correction algorithm and go to S108. S105: Judgment Yes, then （7） If not, proceed to step S107; In the formula, , , , , , and The first The reservoir's initial capacity, final capacity, discharge, inflow, diversion flow, ecological flow, and duration for each time period; S106: Judgment Yes, then If not, execute the backreference traffic correction algorithm and proceed to the next step. （8） And proceed to the next step; S107: The Initial water level of the reservoir during the period , And transfer to S104; S108: Calculate the reservoir's final storage capacity for all time periods using the water level-storage capacity relationship curve. Convert the water level to the final water level of the reservoir for all time periods, and then end the process.

4. The nested long-term, medium-term, and short-term cascade hydropower station forecasting and dispatching system according to claim 3, characterized in that: The reverse reference traffic correction algorithm includes the following steps: S111: Initialization ; S112: Judgment If yes, proceed to the next step; otherwise, go to S117. S113: Judgment If yes, then （9） If not, proceed to step S117; In the formula, , , , , , and The first The reservoir's initial capacity, final capacity, discharge, inflow, diversion flow, ecological flow, and duration for each time period; S114: Judgment If yes, then calculate the initial water level of the reservoir for the first time period using the following formula. , （10） Proceed to the next step; otherwise, proceed to S116. S115: Judgment If yes, then （11） If not, an error occurs, the reference flow cannot be corrected, and the process is redirected to S117. S116： （12） And transfer to S112; In the formula, Indicates the first The water level of the reservoir at the end of the specified time period; S117: Backreference traffic correction terminated, process ended.

5. The nested long-term, medium-term, and short-term cascade hydropower station forecasting and dispatching system according to claim 1, characterized in that: The daily water inflow LSTM prediction module, driven by multidimensional feature data, is executed once at the beginning of each month.

6. The nested long-term, medium-term, and short-term cascade hydropower station forecasting and dispatching system according to claim 1, characterized in that: The hourly water inflow prediction module is executed once a day.

Images