Hybrid pumped storage scheduling method for reducing cascade hydropower operation water consumption rate

Abstract

The invention discloses a hybrid pumped storage scheduling method for reducing the operation water consumption rate of cascade hydropower stations, which comprises the following steps: establishing a cascade water-light-storage complementary power generation scheduling model, and constructing a target function with the minimum total water consumption of the cascade hydropower stations based on the cascade hydropower scheduling model; determining an objective function constraint condition according to the cascade hydropower dispatching model; establishing a solving model on the basis of an Actor-Critic algorithm; time sequence data in the historical data set is used as environment state parameters to be input into the deep reinforcement learning algorithm model, and an optimal day-ahead scheduling strategy is solved through an iterative training network model; and calculating the water consumption rate by using the obtained optimal scheduling strategy result, and evaluating the result. By introducing an Actor-Critic algorithm in deep reinforcement learning, an optimization decision strategy in a complex environment is realized, the economic operation efficiency and the power generation benefit of the system are effectively improved, and the water consumption rate of cascade hydropower is remarkably reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of reducing water consumption rate of hydropower, and in particular to a hybrid pumped storage scheduling method for reducing water consumption rate of cascade hydropower operation. Background Art

[0002] As the most mature, cost-effective, and large-scale development-capable green, low-carbon, clean, and flexible regulating power source in the current power system, pumped storage plays an important role in promoting new energy consumption, improving hydropower peak-shaving and frequency-regulating capabilities, and reducing water consumption rate due to its good coordination effect with wind power, solar power, nuclear power, and thermal power. The water consumption rate is the amount of water consumed by a hydropower station for each unit of electricity generated, and is an important indicator for measuring the economic operation of a hydropower station. The size of the water consumption rate is determined by two factors: the generating head and the efficiency of the turbine generator set. In current research, experts and scholars mainly focus on the study of hydropower complementarity. By constructing a hydropower-photovoltaic complementary system scheduling model and analyzing the comprehensive impact of hydropower-photovoltaic complementarity on hydropower stations, the rationality and effectiveness of the hydropower-photovoltaic complementary mechanism have been verified. However, most of these studies consider a single flexible power source for coordination, which leads to deviations in the execution of plans by power stations.

[0003] The increase in pumped storage units and the output of new energy have a certain impact on the head of cascade power stations, which in turn adjusts the calculation of water consumption rate. However, since it consists of multiple hydropower stations, the scheduling of cascade hydropower requires coordination of water flow and power generation plans between various power stations. The reservoir capacity, water level, water flow, etc. of different hydropower stations will affect the scheduling decision. The scheduling process is complex and requires global consideration and coordinated arrangements to ensure the stable operation of the system and power generation efficiency. Therefore, a suitable scheduling method is needed to deal with complex environments and optimize decision-making strategies.

[0004] Deep reinforcement learning combines the representation learning ability of deep neural networks with the decision optimization ability of reinforcement learning. It can learn high-level abstract features from raw input data and optimize decision strategies by interacting with the environment. The Actor-Critic algorithm is a reinforcement learning algorithm that combines two major components: the policy gradient method (Actor) and the value function approximation (Critic). The Actor is responsible for selecting the appropriate action based on the current state, that is, learning the optimal control strategy; while the Critic is responsible for evaluating the value function of the state and action, that is, judging the quality of the current state or action. Through the interactive learning of these two components, the Actor-Critic algorithm can efficiently handle continuous action spaces and complex environments, and gradually learn the optimal strategy. Summary of the invention

[0005] In order to further study the impact of water-photovoltaic complementarity on water consumption rate, this patent proposes a cascade hydropower pumped storage photovoltaic complementary model. This model is based on the comprehensive power generation system of cascade hydropower stations, including the upgrade of pumped storage units and the supporting output of new energy such as photovoltaics, while taking into account the capacity limitations of regional external transmission channels and the overall inflow of the basin.

[0006] The present invention provides the following technical solution: a hybrid pumped storage scheduling method for reducing the water consumption rate of cascade hydropower operation, comprising:

[0007] Collect the operating parameters of the hydropower station, establish a cascade hydropower-solar-storage complementary power generation scheduling model, and construct an objective function based on the cascade hydropower scheduling model to minimize the total water consumption of the cascade hydropower stations;

[0008] Determine the objective function constraints based on the cascade hydropower dispatch model;

[0009] Establish a solution model based on the Actor-Critic algorithm;

[0010] The time series data in the historical data dataset is input into the deep reinforcement learning algorithm model as the environmental state parameters, and the optimal day-ahead scheduling strategy is solved by iteratively training the network model;

[0011] The water consumption rate is calculated using the optimal scheduling strategy results and the results are evaluated.

[0012] As a preferred solution of the hybrid pumped storage scheduling method for reducing the water consumption rate of cascade hydropower operation described in the present invention, the patent takes the minimum total water consumption of cascade hydropower stations within the scheduling period (one day, 96 points) as the goal to maximize the utilization of water resources. At the same time, in order to avoid frequent and large fluctuations in the output of cascade hydropower stations, a power change penalty term is added. The calculation formula of the objective function is as follows:

[0013]

[0014] Among them, Q i,t is the outflow flow of power station i during period t, m 3 / s; T is the total number of time periods; m is the total number of power stations, Δt is the time period step, s; P i,t is the output of hydropower station i in time period t, MW; θ is the penalty coefficient, which is generally 0.005 according to experience.

[0015] As a preferred solution of the hybrid pumped storage scheduling method for reducing the water consumption rate of cascade hydropower operation described in the present invention, wherein: the objective function constraint conditions include power grid balance constraint, hydropower generator constraint, and reservoir water level constraint;

[0016] The calculation formula of the power grid balance constraint is as follows:

[0017]

[0018] Where, PM t The output of the photovoltaic power station in time period t, WM. D t WM is the section dispatching plan output or local load demand issued by the power grid to the cascaded hydro-PV-storage combined power generation system when connected to the grid during time period t.

[0019] The calculation formula of the hydroelectric generator constraint is as follows:

[0020]

[0021] Where N is the average output of the time period; Q is the power flow; H is the average gross head of the time period. The turbines and pumps in the cascade hydropower station satisfy different NHQ curves, that is, the output of the power station is a binary function of the power flow and gross head, and η is only related to the characteristics of the unit itself.

[0022]

[0023]

[0024] The gross water head is determined by the upper reservoir water level Z up and the water level of the lower reservoir Z down Determine, and the lower reservoir water level is a function of the power generation flow Q.

[0025] The calculation formula of the reservoir water level constraint is as follows:

[0026] V i,t-1 =3600(I i,t -Q i,t )Δt

[0027] I i,t =Q i-1,t-τ +R i,t

[0028]

[0029] Where V i,t is the storage capacity of power station i at the end of time period t; I i,t is the inflow of power station t in period t; t is the flow of power station i-1 in period t-τ after the water lag between power station i and its upstream power station i-1; R i,t is the water flow between power station i-1 and power station i in time period t; and are the power generation flow and abandoned water flow of power station i in time period t respectively.

[0030] V i,t =f izv (Z i,t )

[0031] Z d ≤Z i,t ≤Z u

[0032] The relationship between reservoir water level and storage capacity is nonlinear, and the reservoir water level has upper and lower water level limits.

[0033] As a preferred solution of the hybrid pumped storage scheduling method for reducing the water consumption rate of cascade hydropower operation described in the present invention, wherein: the solution model includes a state space, an action space and a reward mechanism;

[0034] The state space is the state at time t, and the calculation formula is as follows:

[0035] s t ={D t ,Z i,t ,V i,t ,Q p i,t ,t}

[0036] Among them, D t is the section dispatching plan output issued by the power grid to the cascaded hydro-PV combined power generation system when the grid is connected during time period t, Z t is the reservoir water level, V i,t is the storage capacity of power station i at the end of period t, is the power generation flow in time period t, t is time;

[0037] The action space is the action output by the Agent under the state, and the calculation formula is as follows:

[0038] a t = {P 1,i ,P 2,j ,PM t}

[0039] Among them, P 1,i is the output of the traditional water turbine unit, P 2,j The output action of the variable speed pumped storage unit, PM t The output of the photovoltaic power station in time period t;

[0040] The reward mechanism is a reward function r t , the calculation formula is as follows:

[0041] r t =F

[0042] Among them, r t is the reward function.

[0043] Then solve the optimal day-ahead scheduling strategy:

[0044] Set the number of training rounds based on historical data, set one hour as the time step for the day-ahead scheduling task, and set 96 training time steps as one training round;

[0045] Initialize the strategy network parameters θ and value network parameters ω;

[0046] With the current policy π θ Sampling trajectory {s1,a1,r1,s2,a2,r2,...};

[0047] Calculate for each step of data: δ t =r t +γV ω (s t+1 )-V ω (s t );

[0048] Update value parameters

[0049] Update policy parameters

[0050] Repeat the above steps until all rounds are completed.

[0051] In the formula, r t is the instant reward at time t, λ is the discount factor, ▽ θ logπ θ (a t |s t ) indicates to find the gradient of Actor network, ▽ ω V ω (s t ) represents the gradient of the Critic network, α θ It is a soft update parameter.

[0052] As a preferred solution of the hybrid pumped storage scheduling method for reducing the water consumption rate of cascade hydropower operation described in the present invention, wherein: the optimal solution is obtained by the above method, the hydropower output and reservoir status information are obtained, and the water consumption rate is calculated:

[0053] Starting from the total power generation of the cascade power station group, the water consumption rate is calculated based on the power generation of the power station group:

[0054]

[0055] Where λ is the water consumption rate, V is the water volume, and E is the power generation.

[0056] In the second aspect, this patent provides a hybrid pumped storage scheduling method and system for reducing the water consumption rate of cascade hydropower operation, including: a model building module, a target constraint module, a model solving module, a training optimization module and a water consumption rate evaluation and analysis module;

[0057] The model building module is used to establish a cascade hydropower-solar-storage complementary power generation scheduling model, and to construct an objective function based on the cascade hydropower scheduling model to minimize the total water consumption of cascade hydropower stations;

[0058] The target constraint module is used to determine the target function constraint conditions according to the day-ahead dispatch model of cascade hydropower;

[0059] The solution model module is used to establish a solution model based on the Actor-Critic algorithm;

[0060] The training optimization module is used to input the time series data in the historical data dataset as the environmental state parameters into the deep reinforcement learning algorithm model, and solve the optimal day-ahead scheduling strategy by iteratively training the network model;

[0061] The water consumption rate assessment and analysis module obtains the hydropower output and reservoir status information, calculates the water consumption rate, and analyzes and evaluates it.

[0062] In a third aspect, an embodiment of the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, a hybrid pumped storage scheduling method for reducing the water consumption rate of cascade hydropower operation as in the first aspect of the present invention is implemented.

[0063] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, a hybrid pumped storage scheduling method for reducing the water consumption rate of cascade hydropower operation as in the first aspect of the present invention is implemented.

[0064] The beneficial effects of the present invention are as follows: the present invention proposes a hybrid pumped storage scheduling method and system for reducing the water consumption rate of cascade hydropower operation. The method innovatively combines the complementary advantages of cascade hydropower stations, pumped storage units and photovoltaic new energy, and takes into account the external transmission restrictions of the region and the overall inflow of the basin. By introducing the Actor-Critic algorithm in deep reinforcement learning, an optimization decision-making strategy in a complex environment is realized, which effectively improves the economic operation efficiency and power generation efficiency of the system, significantly reduces the water consumption rate of cascade hydropower, and demonstrates the important role of the water-photovoltaic complementary mechanism in improving the utilization of new energy, optimizing water resources management and promoting the development of green and low-carbon electricity. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0066] Figure 1 An overall flow chart of a hybrid pumped storage scheduling method for reducing the water consumption rate of cascade hydropower operation provided by the first embodiment of the present invention;

[0067] Figure 2 Actor-Critic algorithm structure diagram in a hybrid pumped storage scheduling method for reducing water consumption rate of cascade hydropower operation provided by the first embodiment of the present invention;

[0068] Figure 3 A diagram of dispatching results without pumped storage units in a hybrid pumped storage dispatching method for reducing the water consumption rate of cascade hydropower operation provided by the second embodiment of the present invention;

[0069] Figure 4 A diagram of the scheduling results of pumped-storage units included in a hybrid pumped-storage energy scheduling method for reducing the water consumption rate of cascade hydropower operations provided in the second embodiment of the present invention. DETAILED DESCRIPTION

[0070] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0071] Example 1, reference Figure 1 , 2 , which is an embodiment of the present invention, provides a hybrid pumped storage scheduling method for reducing the water consumption rate of cascade hydropower operation, including:

[0072] S1: Establish a cascade hydropower-solar-storage complementary power generation scheduling model, and construct an objective function based on the cascade hydropower scheduling model to minimize the total water consumption of cascade hydropower stations.

[0073] S1.1: The objective function is calculated as follows:

[0074]

[0075] Among them, Q i,t is the outflow flow of power station i during period t, m 3 / s; T is the total number of time periods; m is the total number of power stations, Δt is the time period step, s; P i,t is the output of hydropower station i in time period t, MW; θ is the penalty coefficient, which is generally 0.005 according to experience.

[0076] S2: Determine the objective function constraints based on the day-ahead dispatch model of cascade hydropower.

[0077] S2.1: The objective function constraints include grid balance constraints, hydroelectric generator constraints, and reservoir water level constraints.

[0078] S2.1: The calculation formula of the power grid balance constraint is as follows:

[0079]

[0080] Where, PM t The output of the photovoltaic power station in time period t, WM. D t WM is the section dispatching plan output or local load demand issued by the power grid to the cascaded hydro-PV-storage combined power generation system when connected to the grid during time period t.

[0081] The calculation formula of the hydroelectric generator constraint is as follows:

[0082]

[0083] Where N is the average output of the time period; Q is the power flow; H is the average gross head of the time period. The turbines and pumps in the cascade hydropower station satisfy different NHQ curves, that is, the output of the power station is a binary function of the power flow and gross head, and η is only related to the characteristics of the unit itself.

[0084]

[0085]

[0086] The gross water head is determined by the upper reservoir water level Z up and the water level of the lower reservoir Z down Determine, and the lower reservoir water level is a function of the power generation flow Q.

[0087] The calculation formula of the reservoir water level constraint is as follows:

[0088] V i,t-1 =3600(I i,t -Q i,t )Δt

[0089] I i,t =Q i-1,t-τ +R i,t

[0090]

[0091] Where V i,t is the storage capacity of power station i at the end of time period t; I i,t is the inflow of power station t in period t; t is the flow of power station i-1 in period t-τ after the water lag between power station i and its upstream power station i-1; R i,t is the water flow between power station i-1 and power station i in time period t; and are the power generation flow and abandoned water flow of power station i in time period t respectively.

[0092] V i,t =f i zv (Z i,t )

[0093] Z d ≤Z i,t ≤Z u

[0094] The relationship between reservoir water level and storage capacity is nonlinear, and the reservoir water level has upper and lower water level limits.

[0095] S3: Establish a solution model based on the Actor-Critic algorithm.

[0096] S3.1: The solution model includes state space, action space and reward mechanism;

[0097] S3.2: The state space is the state at time t, and the calculation formula is as follows:

[0098] s t ={D t ,Z i,t ,V i,t ,Q p i,t ,t}

[0099] Among them, D t is the section dispatching plan output issued by the power grid to the cascaded hydro-PV combined power generation system when the grid is connected during time period t, Z t is the reservoir water level, V i,t is the storage capacity of power station i at the end of period t, is the power generation flow in time period t, and t is time.

[0100] The action space is the action output by the Agent under the state, and the calculation formula is as follows:

[0101] a t = {P 1,i ,P 2,j ,PM t}

[0102] Among them, P 1,i is the output of the traditional turbine unit, P 2,j The output action of the variable speed pumped storage unit, PM t The output of the photovoltaic power station in time period t.

[0103] The reward mechanism is a reward function r t , the calculation formula is as follows:

[0104] r t =-F

[0105] Among them, r t is the reward function.

[0106] S4: Input historical data as environmental state parameters into the Actor-Critic algorithm model, and solve the optimal day-ahead scheduling strategy by iteratively training the network model.

[0107] S4.1: Set the number of training rounds according to historical data, set one hour as the time step of the day-ahead scheduling task, and set 96 training time steps as one training round.

[0108] S4.2: The specific training process is as follows:

[0109] Initialize the policy network parameters θ and the value network parameters ω.

[0110] With the current policy π θ Sampling trajectory {s1,a1,r1,s2,a2,r2,…}.

[0111] For each step of data, calculate:

[0112] Update value parameters

[0113] Update policy parameters

[0114] Repeat the above steps until all rounds are completed.

[0115] In the formula, r t is the instant reward at time t, λ is the discount factor, ▽ θ logπ θ (a t |s t ) indicates to find the gradient of the Actor network, ▽ ω V ω (s t ) represents the gradient of the Critic network, α θ It is a soft update parameter.

[0116] S5: Obtain the optimal solution through the above method, obtain the hydropower output and reservoir status information, and calculate the water consumption rate:

[0117] S5.1: Starting from the total power generation of the cascade power station group, the water consumption rate is calculated based on the power generation of the power station group.

[0118]

[0119] Where λ is the water consumption rate, V is the water volume, and E is the power generation.

[0120] S5.2: Evaluate and analyze the results.

[0121] If the above functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0122] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0123] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0124] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0125] Example 2, reference Figure 3 , 4 , which is an embodiment of the present invention, provides a hybrid pumped storage scheduling method for reducing the water consumption rate of cascade hydropower operation. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0126] The four cascade power stations in the Beipan River Basin are taken as the research objects. From the main power station to the tail power station, they are Shannipo, Guangzhao, Mamaya and Dongqing (replaced by No. 1, No. 2, No. 3 and No. 4 respectively). Among them, No. 4 power station has added a pumped storage unit on the original basis, so the other three power stations can be regarded as an overall power generation system. No. 4 power station is an output system combined with photovoltaic power generation to ensure the full absorption of photovoltaic energy. The basic information and constraints of the power station are shown in Tables 1, 2 and 3.

[0127] Table 1: Basic data of the unit

[0128] unit No. 1 No. 2 No.3 No. 4 Adjustment capability sky sky sky sky Number of units 2 4 3 4 Unit capacity(MW) 92.31 265.3 183.65 224.5 Oscillation area (MW) 0,60 0,70 0,50 100,120 Water flow stagnation time (h) 0 2 1 1 Water level range(m) 865,885 691,745 580,585 483,490

[0129] Table 2: Parameters of No. 4 pumped storage unit

[0130] Main parameters Installed capacity (MW) 200 Number of units 4 Rated output(MW) 204.08

[0131] Table 3: Parameters of pump turbines No. 1, 2 and 3

[0132]

[0133] Collect historical hydrological data, power load data, wind power and photovoltaic power generation data in the basin, etc. These data cover a whole year, with hourly time steps, and contain a total of 8760 time steps.

[0134] A cascade hydropower-solar-storage complementary power generation scheduling model is established, and an objective function with the minimum total water consumption of cascade hydropower stations is constructed based on the cascade hydropower scheduling model.

[0135] Determine the objective function constraints based on the cascade hydropower scheduling model.

[0136] A solution model is established based on the Actor-Critic algorithm.

[0137] The time series data in the historical data dataset is input into the deep reinforcement learning algorithm model as the environmental state parameters, and the optimal day-ahead scheduling strategy is solved by iteratively training the network model.

[0138] The number of training rounds is set to 500, and each round contains 96 time steps. The model parameters are updated every hour, and the scheduling strategy is continuously adjusted through the reinforcement learning algorithm until convergence reaches the preset accuracy.

[0139] The water consumption rate is calculated using the optimal scheduling strategy results and the results are evaluated.

[0140] In the case of grid connection, the output of the segmented dispatch plan or local load demand issued by the power grid to the cascaded hydro-solar energy storage combined power generation system only considers historical data, selects the classic load scenario of two peaks and one valley (morning peak and evening peak), and finally obtains two dispatch results with and without the participation of pumped storage units, such as Figure 3 and Figure 4 shown.

[0141] from Figure 3 and Figure 4 The empirical analysis results clearly show that after the pumped storage unit was integrated into the system, the pumped storage unit showed a significant effect of giving priority to photovoltaic power consumption, provided that the cross-section load remained stable and the photovoltaic output power did not change. This effect is reflected in that, compared with before the pumped storage unit was added, when photovoltaic power was in the power generation state, except for the GM hydropower station, the comprehensive output power of the other three hydropower stations was improved with the assistance of the pumped storage unit, thus effectively maintaining the balance of the system load.

[0142] Table 4: Water consumption results

[0143]

[0144]

[0145] As can be seen from Table 4, the addition of pumped storage units has had a significant positive impact on the overall water consumption rate of cascade power stations in the basin. Taking the Beipanjiang River Basin as an example, by introducing the No. 3 pumped storage power station, the water consumption rate of cascade hydropower in the basin has been significantly reduced by 21.8%. This achievement is not only reflected in the optimization of data, but also actually ensures the effective absorption of photovoltaic output, further improving the comprehensive performance of the power system in the basin.

[0146] An in-depth analysis of the fundamental reason for the reduction in water consumption after the addition of pumped storage units can be attributed to the efficient use of photovoltaic energy storage water. During the pumped storage process, when the photovoltaic output is high, the excess electricity is used to pump water from a low water level to a high water level for storage. Later, when power generation is needed, the stored water is released and converted into hydropower, thus achieving the repeated use of water. This process effectively reduces water consumption while maintaining the overall power generation demand unchanged, thereby reducing the water consumption rate.

[0147] From a more macro perspective, this transformation can also be seen as a form conversion and space-time optimization of energy. The original way of directly participating in meeting the load demand with light energy has been transformed into a way of converting light energy into water energy through pumped storage units, and then indirectly meeting the load demand with water power. This transformation brings many benefits: first, it transforms the originally uncertain photovoltaic output into more controllable hydropower generation, thereby reducing the need for accurate prediction of photovoltaic output and reducing the potential risks caused by prediction errors; second, the introduction of pumped storage units enhances the regulation capacity of the power system, which helps to maintain the stability of the power grid when the load fluctuates; finally, from an economic perspective, by optimizing the use of water resources, unnecessary water abandonment is reduced, the utilization efficiency of water resources is improved, and higher economic benefits are brought to the cascade power stations in the basin.

[0148] In summary, the addition of pumped storage units has shown significant advantages in reducing the water consumption rate of cascade power stations in the basin, promoting the absorption of photovoltaic output, enhancing the stability of the power system and improving economic benefits, providing strong technical support for the development of green and low-carbon electricity.

[0149] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A hybrid pumped storage scheduling method for reducing the water consumption rate of cascade hydropower operation, characterized in that: include: Collect the operating parameters of the hydropower station, establish a cascade hydropower-solar-storage complementary power generation scheduling model, and construct an objective function based on the cascade hydropower scheduling model to minimize the total water consumption of the cascade hydropower stations; Determine the objective function constraints based on the cascade hydropower dispatch model; Establish a solution model based on the Actor-Critic algorithm; The time series data in the historical data dataset is input into the deep reinforcement learning algorithm model as the environmental state parameters, and the optimal day-ahead scheduling strategy is solved by iteratively training the network model; The water consumption rate is calculated using the optimal scheduling strategy results and the results are evaluated.

2. The hybrid pumped storage scheduling method for reducing the water consumption rate of cascade hydropower operation according to claim 1, characterized in that: The calculation formula of the objective function is as follows: Among them, Q i,t is the outflow of power station i in period t; T is the total number of periods; m is the total number of power stations, Δt is the period step length; P i,t is the output of hydropower station i in time period t; θ is the penalty coefficient.

3. The hybrid pumped storage scheduling method for reducing the water consumption rate of cascade hydropower operation as claimed in claim 2 is characterized by: The objective function constraints include power grid balance constraints, hydroelectric generator constraints, and reservoir water level constraints.

4. The hybrid pumped storage scheduling method for reducing the water consumption rate of cascade hydropower operation as claimed in claim 3 is characterized by: The calculation formula of the power grid balance constraint is as follows: Among them, PM t is the output of the photovoltaic power station in time period t, D t It is the section dispatching plan output or local load demand issued by the power grid to the cascaded hydro-PV-storage combined power generation system when the system is connected to the grid during time period t; The calculation formula of the hydroelectric generator constraint is as follows: Among them, N is the average output of the time period; Q is the power flow; H is the average gross head of the time period; the turbines and pumps in the cascade hydropower station meet different NHQ curves respectively, and the output of the power station is a binary function of the power flow and gross head; η is the adjustment coefficient, which is only related to the characteristics of the unit itself; The gross water head is determined by the upper reservoir water level Z up and the water level of the lower reservoir Z down Determine, and the lower reservoir water level is a function of the power generation flow Q; The calculation formula of the reservoir water level constraint is as follows: V i,t-1 =3600(I i,t -Q i,t )Δt I i,t =Q i-1,t-τ +R i,t Where V i,t is the storage capacity of power station i at the end of time period t; I i,t is the inflow of power station t in period t; t is the flow of power station i-1 in period t-τ after the water lag between power station i and its upstream power station i-1; R i,t is the water flow between power station i-1 and power station i in time period t; and are the power generation flow and abandoned water flow of power station i in time period t respectively; In i,t =f i zv (With i,t ) WITH d ≤Z i,t ≤Z u The relationship between reservoir water level and storage capacity is nonlinear, and the reservoir water level has upper and lower water level limits.

5. The hybrid pumped storage scheduling method for reducing the water consumption rate of cascade hydropower operation as claimed in claim 4, characterized in that: The solution model includes a state space, an action space and a reward mechanism; The solution model includes a state space, an action space and a reward mechanism; The state space is the state at time t, and the calculation formula is as follows: s t ={D t ,Z i,t ,V i,t ,Q p i,t ,t} Among them, D t is the section dispatching plan output issued by the power grid to the cascaded hydro-PV combined power generation system when the grid is connected during time period t, Z t is the reservoir water level, V i,t is the storage capacity of power station i at the end of period t, is the power generation flow in time period t, t is time; The action space is the action output by the Agent under the state, and the calculation formula is as follows: to t ={P 1,i ,P 2,j ,P.M t } Among them, P 1,i is the output of the traditional turbine unit, P 2,j The output action of the variable speed pumped storage unit, PM t The output of the photovoltaic power station in time period t; The reward mechanism is a reward function r t , the calculation formula is as follows: r t ＝-F Among them, r t is the reward function.

6. The hybrid pumped storage scheduling method for reducing the water consumption rate of cascade hydropower operation as claimed in claim 5, characterized in that: The Actor-Critic algorithm solves the optimal day-ahead scheduling strategy by: Set the number of training rounds based on historical data, set one hour as the time step for the day-ahead scheduling task, and set 96 training time steps as one training round; Initialize the policy network parameters θ and value network parameters ω; With the current policy π θ Sampling trajectory {s1,a1,r1,s2,a2,r2,...}; For each step of data, calculate: Update value parameters Update policy parameters Repeat the above steps until all rounds are completed. In the formula, r t is the instant reward at time t, λ is the discount factor, Indicates the gradient of the Actor network. Indicates the gradient of the Critic network, α θ It is a soft update parameter.

7. The hybrid pumped storage scheduling method for reducing the water consumption rate of cascade hydropower operation as claimed in claim 6, characterized in that: Find the optimal solution, obtain hydropower output and reservoir status information, and calculate water consumption rate; Starting from the total power generation of the cascade power station group, the water consumption rate is calculated based on the power generation of the power station group: Where λ is the water consumption rate, V is the water volume, and E is the power generation.

8. A hybrid pumped storage dispatching system for reducing the water consumption rate of cascade hydropower operation by using the method as claimed in any one of claims 1 to 7, characterized in that: Model building module, target constraint module, model solving module, training optimization module and water consumption rate evaluation and analysis module; The model building module is used to establish a cascade hydro-photovoltaic storage complementary power generation scheduling model, and to construct an objective function based on the cascade hydropower scheduling model to minimize the total water consumption of cascade hydropower stations; The target constraint module is used to determine the target function constraint conditions according to the cascade hydropower day-ahead dispatch model; The solution model module is used to establish a solution model based on the Actor-Critic algorithm; The training optimization module is used to input the time series data in the historical data data set as the environmental state parameter into the deep reinforcement learning algorithm model, and solve the optimal day-ahead scheduling strategy by iteratively training the network model; The water consumption rate evaluation and analysis module obtains hydropower output and reservoir status information, calculates the water consumption rate, and performs analysis and evaluation thereof.

9. A computer device comprising: A memory and a processor; the memory stores a computer program, wherein the processor implements the steps of any method as claimed in claim 1 when executing the computer program.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images