Pumped storage unit participation real-time optimization scheduling method and system

Abstract

The invention discloses a method and system for a pumped storage unit to participate in real-time optimization scheduling, and the method comprises the steps: expanding a real-time optimization time period into a standard time period and an expansion time period, and carrying out the electricity price prediction through a neural network, and obtaining a system electricity price; in the target function, a pumped storage unit power generation cost item is added in the standard time period, and a pumped storage unit power generation cost item considering electricity price prediction is added in the extended time period; in the constraint conditions, continuous variable constraint, integer variable constraint and energy storage state constraint of the pumped storage unit are added in the standard optimization time period, and continuous variable constraint and energy storage state constraint of the pumped storage unit are added in the expansion time period; calculating a pumped storage unit output plan in real time; the method can improve the enthusiasm of pumped storage to participate in the market, stimulate the quick response effect, guide the pumped storage unit to actively participate in peak load shifting, and contribute to improving the balance capacity of a power grid, reducing the power generation cost and improving the economic benefit of a pumped storage plant station.

Description

Technical Field

[0001] This invention relates to scheduling methods and systems, and more particularly to a method and system for real-time optimized scheduling of pumped storage units. Background Technology

[0002] With the continuous advancement of the construction of new power systems, the proportion of new energy sources such as wind and solar power in the power supply is gradually increasing. However, due to their randomness and volatility, these new energy sources pose significant challenges to the operation of the power system, increasing the pressure on the power grid to ensure supply and promote consumption.

[0003] Against this backdrop, pumped-storage hydroelectric power plants have become an important tool for improving grid flexibility due to their ability to flexibly regulate power system consumption. Furthermore, by storing and releasing energy within the power system, pumped-storage power plants provide the grid with unique regulation capabilities. This flexibility is particularly important in the current system, as the share of intermittent renewable resources continues to increase.

[0004] However, how to consistently utilize this flexibility in a real-time optimized scheduling environment to minimize power generation costs remains an unresolved issue. Many existing methods, when dealing with the flexible scheduling of pumped storage power plants, have failed to adequately consider optimizing their operating strategies in a dynamically changing market environment, thus preventing the power plants from fully realizing their regulatory functions and consequently affecting their overall efficiency. Summary of the Invention

[0005] Purpose of the invention: The purpose of this invention is to provide a method for pumped storage units to participate in real-time optimized scheduling to solve the problems in the process of pumped storage power stations participating in real-time scheduling, guide pumped storage units to actively participate in peak shaving and valley filling, thereby improving the grid balance capability, reducing power generation costs, and improving the economic benefits of pumped storage power plants.

[0006] Technical solution: The optimized scheduling method of the present invention includes the following steps:

[0007] The real-time optimization period is expanded to include standard and extended periods, and a fully connected neural network model is used to predict the electricity price during the extended period.

[0008] Considering the energy storage boundary of a pumped storage power station, objective functions and constraints for the standard time period and extended time period of the real-time optimization model are constructed respectively.

[0009] In the objective function, compared with the traditional real-time optimization scheduling without pumped storage participating in the market, the standard time period adds a pumped storage unit power generation cost term, and the extended time period adds a consideration of the pumped storage unit power generation cost based on electricity price.

[0010] Compared to traditional real-time optimization scheduling without pumped storage participating in the market, the standard time period adopts continuous variable constraints, integer variable constraints, energy storage state constraints and standard real-time optimization constraints for pumped storage units, while the extended time period adopts continuous variable constraints and energy storage state constraints for pumped storage units.

[0011] Optimize the calculation output of real-time unit output plan, pumped storage unit status and output;

[0012] The system publishes real-time optimized dispatch results that take electricity prices into account, including power generation plans for conventional units and pumped-storage power generation plans for pumped-storage units.

[0013] Preferably, the constraints of the extended time period include:

[0014]

[0015] in, and This indicates the lower and upper limits of the pumping output of a pumped storage unit. and This indicates the lower and upper limits of the power generation output of pumped storage units; This represents the amount of water stored in a pumped storage power station during time period t, corresponding to the amount of electricity stored. and This represents the lower and upper limits of the power plant's energy storage. Considering the timeliness requirements of real-time optimization calculations, the optimization training variables and constraints within the extended time period only consider continuous variables and SOC constraints, and do not consider non-continuous Boolean variables.

[0016] Preferably, the constraints for the standard time period include system traditional balance constraints, pumped storage unit state and transition logic constraints, and continuous variables and SOC constraints of the pumped storage unit.

[0017] Preferably, the system's conventional equilibrium constraints are as follows:

[0018]

[0019] Among them, D t For the system net load, G psh A collection of pumped storage units. The pumping output of the unit during time period t is g. The power output of the unit during time period t is the power generation capacity of the unit in time period g.

[0020] The state and transition logic constraints of the pumped storage unit are as follows:

[0021]

[0022] Among them, M g Let [off, gen, pump] be the set of pumped-storage unit states. Let m be a feasible state that can be transitioned to. The variable is 0-1, representing the state m of the pumped storage unit during time period g in time t. The variable is 0-1, representing the state of the pumped storage unit g changing from state m to state n within time period t.

[0023] Preferably, the continuous variables and SOC constraints of the pumped storage unit are as follows:

[0024]

[0025]

[0026] in, and This indicates the lower and upper limits of the pumping output of a pumped storage unit. and This indicates the lower and upper limits of the power generation output of pumped storage units; This represents the amount of water stored in a pumped storage power station during time period t, corresponding to the amount of electricity stored. and This indicates the lower and upper limits of energy storage for power plants.

[0027] Preferably, the electricity price for the extended period is predicted using a fully connected neural network model, and the optimization objective formula for the extended period is as follows:

[0028]

[0029] Specifically, in the period t0 before the first period t1, the electricity price prediction for period t is obtained by learning the electricity price prediction through a fully connected neural network. end Electricity price forecast results from +1 to the final time period T Calculate the pumping cost and power generation revenue of pumped storage unit g at time t.

[0030] The optimized scheduling system of the present invention includes:

[0031] The extended-period electricity price prediction module is used to predict the extended-period electricity price through a neural network, providing electricity price prediction input for the real-time optimization of the extended-period module.

[0032] The real-time optimization extended time period module is used to extend the real-time optimization time period into a standard time period and an extended time period, and to construct the objective function and constraints of the real-time optimization model for the extended time period.

[0033] Real-time optimization standard time period module: used to consider the energy storage boundary of the pumped storage power station in real time, and to construct the objective function and constraints of the real-time optimization model standard time period;

[0034] The real-time optimization scheduling calculation module is used to combine the objective functions and constraints of the standard time period and the extended time period to construct a real-time optimization scheduling model, solve the real-time optimization scheduling model, and obtain the power dispatch operation plan for pumped storage units to participate in real-time optimization scheduling based on electricity price forecasts.

[0035] The real-time optimization results publishing module is used to publish the results of real-time optimization scheduling, including the power generation plan of traditional units and the pumped-storage power generation plan of pumped storage units.

[0036] Preferably, the objective function of the complete real-time optimization scheduling model in real-time optimization across the entire time period is as follows:

[0037]

[0038] Where F is the real-time optimization scheduling target, g is the unit participating in the real-time optimization scheduling, and G is the set of units. psh For the unit assembly, G psh For the set of pumped storage units, t is the calculation period. start To t end The current standard calculation period is defined as real-time calculation, where C is the unit cost function, and p... g,t q g,t u g,t These represent the unit output, pumped storage unit output, and unit binary integer variable, respectively. T is the end of the extended calculation period, and t0 is the value of this calculation based on time period t. start The optimized scheduling execution time for the initial period occurs at time t. start Before, t end +1 to T is the extended time period. The marginal electricity price based on deep learning is executed for time period t0. The pumping output of the unit during time period t is g. The power output of the unit during time period t is the power generation capacity of the unit in time period g.

[0039] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages: by participating in real-time optimization scheduling calculations based on electricity price prediction of pumped storage, it does not require a large amount of manpower, and the calculation speed can meet the needs of practical applications. It effectively solves the problems of traditional multi-resource optimization, which mainly focuses on short-term and periodic connections, relies on experience, and is inefficient. It has broad prospects for promotion. Attached Figure Description

[0040] Figure 1 This is a schematic diagram of the method flow of the present invention;

[0041] Figure 2 This is a schematic diagram of the system structure of the present invention. Detailed Implementation

[0042] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0043] like Figure 1 As shown, the method for real-time optimization scheduling of pumped storage units includes the following steps:

[0044] S1. Expand the real-time optimization period into a standard period and an extended period, considering factors such as weather, historical electricity prices, and date characteristics. Predict the electricity price for the extended real-time optimization period using a fully connected neural network model, and construct the objective function term for the extended period. Calculate the electricity price prediction for the extended period, specifically including:

[0045] set up For extended time period t end +1 to T system electricity price, x t,f f is the input feature required for electricity price forecasting. M For the trained prediction model:

[0046]

[0047] Compared to traditional real-time optimization scheduling that does not participate in the pumped storage market:

[0048] In the objective function, a pumped storage unit power generation cost term is added to the standard time period, and a pumped storage unit power generation cost based on electricity price is added to the extended time period.

[0049] In the constraints, continuous variable constraints, integer variable constraints, energy storage state constraints, and standard real-time optimization constraints are used for pumped-storage units in the standard time period, while continuous variable constraints and energy storage state constraints are used for pumped-storage units in the extended time period.

[0050] S2. Considering the energy storage boundary of the pumped storage power station, construct the objective function and constraints for the standard time period of the real-time optimization model. The objective function is as follows:

[0051]

[0052] Where F1 is the standard time period item of the real-time optimized scheduling target, g is the unit participating in the real-time optimized scheduling, G is the set of units, and t is the calculation time period. start To t end The current standard calculation period is defined as real-time calculation, where C is the unit cost function, and p... g,t q g,t u g,t These represent the unit output, pumped storage unit output, and unit binary integer variables, respectively.

[0053] The constraints for the standard time period specifically include:

[0054] System equilibrium constraints:

[0055]

[0056] Among them, D t For the system net load, G psh A collection of pumped storage units. The pumping output of the unit during time period t is g. The power output of the unit during time period t is the power generation capacity of the unit in time period g.

[0057] Pumped storage unit status and transition logic constraints:

[0058]

[0059] Among them, M g Let [off, gen, pump] be the set of pumped-storage unit states. Let m be a feasible state that can be transitioned to. The variable is 0-1, representing the state m of the pumped storage unit during time period g in time t. The variable is 0-1, representing the state of the pumped storage unit g changing from state m to state n within time period t.

[0060] Continuous variables and SOC constraints for pumped storage units:

[0061]

[0062] in, and This indicates the lower and upper limits of the pumping output of a pumped storage unit. and This indicates the lower and upper limits of the power generation output of pumped storage units; This represents the amount of water stored in a pumped storage power station during time period t, corresponding to the amount of electricity stored. and This represents the lower and upper limits of power plant energy storage. Considering the timeliness requirements of real-time optimization calculations, the optimization training variables and constraints within the extended time period only consider continuous variables and SOC constraints, and do not consider non-continuous Boolean variables.

[0063] In addition to other traditional real-time optimization practical constraints, including conventional unit output constraints and unit ramping constraints.

[0064] S3. Construct the objective function and constraints for the extended time period of the real-time optimization model, as follows:

[0065] The objective function term specifically includes:

[0066]

[0067] Where F2 is the real-time optimization scheduling target extension period term, t is the calculation period, and G psh For pumped storage units, tend +1 to T is the extended time period.

[0068] Standard time period constraints, extended time period constraints for the real-time optimization scheduling model of pumped storage units based on electricity price forecasts:

[0069]

[0070] in, and This indicates the lower and upper limits of the pumping output of a pumped storage unit. and This indicates the lower and upper limits of the power generation output of pumped storage units; This represents the amount of water stored in a pumped storage power station during time period t, corresponding to the amount of electricity stored. and This indicates the lower and upper limits of energy storage for power plants.

[0071] S4. Combine the objective functions and constraints of the extended time period and the standard time period, and then solve the optimization scheduling model to obtain the power dispatch operation plan for pumped storage units participating in real-time optimization based on electricity price forecasts. The optimization scheduling model is as follows:

[0072]

[0073] Where F is the real-time optimization scheduling target, g is the unit participating in the real-time optimization scheduling, and G is the set of units. psh For the unit assembly, G psh For the set of pumped storage units, t is the calculation period. start To t end The current standard calculation period is defined as real-time calculation, where C is the unit cost function, and p... g,t q g,t u g,t These represent the unit output, pumped storage unit output, and unit binary integer variable, respectively. T is the end of the extended calculation period, and t0 is the value of this calculation based on time period t. start The optimized scheduling execution time for the initial period occurs at time t. start Before, t end +1 to T is the extended time period. The marginal electricity price based on deep learning is executed for time period t0. The pumping output of the unit during time period t is g. The power output of the unit during time period t is the power generation capacity of the unit in time period g.

[0074] S5. Use the solver to perform optimization calculations, obtain the calculation results, and then publish the optimized scheduling results of multi-energy complementarity, including the power generation plans of new energy, hydropower, thermal power units and other units, as well as the charging and discharging plans of pumped storage units.

[0075] like Figure 2 As shown, the pumped storage unit participates in the real-time optimization scheduling system, including:

[0076] The real-time optimization extended time period module is used to extend the real-time optimization time period into a standard time period and an extended time period. It considers the real-time unit pumped storage power station energy storage and other general boundaries, and constructs the objective function and constraints of the standard time period of the real-time optimization model.

[0077] The extended time period electricity price prediction module is used to predict electricity prices for extended time periods. It takes into account factors such as weather, historical electricity prices, and date characteristics. It uses a neural network to predict the extended time period electricity price in real time optimization. It also takes into account the energy storage boundary of pumped storage power stations and constructs a real-time optimization model with an extended time period objective function and constraints, providing electricity price prediction input for the real-time optimization extended time period module.

[0078] The real-time optimization standard time period module is used to consider the energy storage boundary of the pumped storage power station in real time, construct the objective function and constraints of the real-time optimization model standard time period, minimize the power generation cost of the unit as the optimization objective, and consider the system balance constraints after the participation of pumped storage units, the power generation and pumping power constraints of pumped storage units, the state constraints of pumped storage units, the energy storage constraints of pumped storage units, the state constraints of pumped storage, and the constraints of non-pumped storage units.

[0079] The real-time optimization scheduling calculation module is used to combine the objective functions and constraints of the standard time period and the extended time period to construct a real-time optimization scheduling model, solve the real-time optimization scheduling model, and obtain the power dispatch operation plan for pumped storage units to participate in real-time optimization scheduling based on electricity price forecasts.

[0080] The real-time optimization results publishing module is used to publish the real-time optimization results of pumped storage units based on electricity price forecasts, including the power generation plans of traditional units and the pumping-power generation plans of pumped storage units.

[0081] Taking the 11:00 real-time optimization plan of the provincial power grid as an example, the modeling and calculation of pumped storage units are considered to carry out optimized scheduling.

[0082] S1. Extended period electricity price forecasting: A 2-hour standard period is used, i.e., 11:00-13:00. The remaining time of the day is the extended period, i.e., 13:15-24:00. Electricity price forecasting is performed for the extended period with a granularity of 15 minutes.

[0083] S2. Based on electricity price forecasts, construct real-time optimization objective functions for the standard and extended time periods. The objective function for the standard time period has a time granularity of 5 minutes and consists of minimizing the generator cost. The objective function for the extended time period has a time granularity of 15 minutes and consists of minimizing the pumped-storage unit's generator cost. Minimizing the generator cost within the extended time period is equivalent to maximizing the pumped-storage unit's revenue, i.e., the pumped-storage unit generates electricity during periods of high electricity prices and pumps water during periods of low electricity prices.

[0084] S3. Construct real-time optimization constraints. Constraints consist of standard time period constraints and extended time period constraints. Within the standard time period, the constraint time granularity is 5 minutes, comprising system balance constraints considering pumped storage units, pumped storage unit power generation and pumping power constraints, pumped storage unit state constraints, pumped storage unit energy storage constraints, pumped storage state constraints, and constraints for non-pumped storage units. Within the extended time period, the constraint time granularity is 15 minutes, comprising pumped storage unit power generation and pumping power constraints, pumped storage unit energy storage constraints, and pumped storage state constraints.

[0085] S4. Perform optimization calculations on the optimization problem to obtain the pumped storage unit status and output at a 5-minute granularity within the standard time period, as well as the output of other units.

[0086] S5. Analyze the results of pumped storage units participating in real-time optimized scheduling. Under the objectives and constraints of the standard time period and the extended time period, pumped storage units play a regulatory role in real-time optimized scheduling, generating electricity during high electricity price periods and pumping water during low electricity price periods. Within the standard time period, pumped storage units participate in real-time optimized scheduling under constraints, playing a regulatory role and improving efficiency.

[0087] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0088] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0089] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0090] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0091] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for participating in real-time optimal dispatching of a pumped storage unit, characterized in that, The method comprises the following steps: The real-time optimization period is extended into a standard period and an extended period, and a full connection neural network model is used to predict the electricity price of the extended period; Considering the energy storage boundary of the pumped storage power station, the objective function and the constraint condition of the real-time optimization model of the standard period and the extended period are constructed respectively; In the objective function, compared with the traditional real-time optimization dispatching without pumped storage participating in the market, the generation cost item of the pumped storage unit is added in the standard period, and the generation cost of the pumped storage unit based on the electricity price is considered in the extended period; In the constraint condition, compared with the traditional real-time optimization dispatching without pumped storage participating in the market, the continuous variable constraint, the integer variable constraint, the energy storage state constraint and the standard real-time optimization constraint of the pumped storage unit are adopted in the standard period, and the continuous variable constraint and the energy storage state constraint of the pumped storage unit are adopted in the extended period; The optimization calculation outputs the real-time unit output plan, the pumped storage unit state and the output; The real-time optimization dispatching result considering the electricity price is released, including the generation plan of the traditional unit and the pumping-generating plan of the pumped storage unit.

2. The scheduling method of claim 1, wherein, The constraint condition of the extended period comprises: wherein, and denote the lower and upper limits of the pumped storage power output of the pumped storage unit, and denote the lower and upper limits of the power output of the pumped storage unit; denote the amount of stored energy corresponding to the amount of stored water of the pumped storage power station at the time period t; and denote the lower and upper limits of the stored energy of the power station.

3. The scheduling method of claim 1, wherein, The constraint condition of the standard period comprises the traditional balance constraint, the pumped storage unit state and the conversion logic constraint, and the pumped storage unit continuous variable and SOC constraint.

4. The scheduling method of claim 3, wherein, The traditional balance constraint is as follows: wherein D t is the system net load, G psh is the set of pumped storage units, is the pumped hydro output of unit g in period t, is the generated hydro output of unit g in period t. The pumped storage unit state and the conversion logic constraint are as follows: where M g is the set of pumped storage unit states [off, gen, pump], is the feasible states that state m can transition to, is a 0-1 variable representing pumped storage unit g in period t in state m, is a 0-1 variable representing the transition from state m to n for pumped storage unit g in period t.

5. The scheduling method of claim 3, wherein, The pumped storage unit continuous variable and SOC constraint are as follows: wherein, and denote the lower and upper limits of the pumped storage power output of the pumped storage unit, and denote the lower and upper limits of the power output of the pumped storage unit; denote the storage capacity corresponding to the power storage capacity of the pumped storage power station at time period t; and denote the lower and upper limits of the storage capacity of the power station.

6. The scheduling method of claim 1, wherein, The formula for constructing the optimization objective function of the extended period by using the electricity price prediction result is as follows: Wherein, at the t0 period before the first period t1, the period t obtained by learning the electricity price prediction through the fully connected neural network end The electricity price prediction result from +1 to the final period T The pumped storage cost and power generation benefit calculation of the pumped storage group g at time t 7. A pumped storage unit participating in a real-time optimal dispatch system, characterized in that, It comprises: The extended period electricity price prediction module is used to predict the extended period electricity price by using the neural network, and provide the electricity price prediction input for the real-time optimization extended period module; The real-time optimization extended period module is used to extend the real-time optimization period into a standard period and an extended period, and construct the real-time optimization model of the extended period objective function and the constraint condition; The real-time optimization standard period module is used to consider the energy storage boundary of the real-time unit pumped storage power station, and construct the real-time optimization model of the standard period objective function and the constraint condition; The real-time optimization dispatching calculation module is used to combine the standard period and the extended period objective function and the constraint condition, construct the real-time optimization dispatching model, solve the real-time optimization dispatching model, and obtain the power dispatching operation plan of the pumped storage unit participating in the real-time optimization dispatching based on the electricity price prediction; The real-time optimization result release module is used to release the result of the real-time optimization dispatching, including the generation plan of the traditional unit and the pumping-generating plan of the pumped storage unit.

8. The dispatch system of claim 7, wherein, The objective function of the complete real-time optimization dispatching model is as follows: Where F is the real-time optimization scheduling target, g is the unit participating in the real-time optimization scheduling, and G is the set of units. psh For the unit assembly, G psh For the set of pumped storage units, t is the calculation period. start To t end Let C be the standard calculation period for real-time calculation, and p be the unit cost function. g,t q g,t u g,t These represent the unit output, pumped storage unit output, and unit binary integer variable, respectively. T is the end of the extended calculation period, and t0 is the value of this calculation based on time period t. start The optimized scheduling execution time for the initial period occurs at time t. start Before, t end +1 to T is the extended time period. The marginal electricity price based on deep learning is applied for time period t0. The pumping output of the unit during time period t is g. The power output of the unit during time period t is the power generation capacity of the unit in time period g.

9. A computer device, comprising: The computer program is executed by the processor to realize the steps of the pumped storage unit participating in the real-time optimization dispatching method in any one of claims 1-6.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to realize the steps of the pumped storage unit participating in the real-time optimization dispatching method in any one of claims 1-6.

Images