Profit distribution method under joint dispatch of multi-agent hybrid pumped storage power stations

Abstract

This invention discloses a method for distributing benefits under the joint scheduling of multi-agent hybrid pumped-storage power stations. First, grid system load data is collected to generate peak and valley electricity prices and daily average electricity prices. Second, a medium- and long-term scheduling model for the hybrid pumped-storage power station is established and solved. Third, a short-term scheduling model for the hybrid pumped-storage power station is established and solved. Finally, based on the Shapley value method, the benefits gained under the joint scheduling of the multi-agent hybrid pumped-storage power station are distributed. This invention fairly distributes the benefits gained from the coordinated operation of the various power generation entities within the hybrid pumped-storage power station, safeguarding the normal operation of the hybrid pumped-storage power station.

Description

Technical Field

[0001] The present invention belongs to the technical field of water conservancy projects, and in particular relates to a method for distributing benefits under the joint dispatch of a multi-agent hybrid pumped-storage power station. Background Art

[0002] Numerous studies, both domestically and internationally, have demonstrated that the integration and transformation of cascade hydropower into hybrid pumped-storage power stations can significantly improve the efficiency and flexibility of cascade systems, and are therefore considered to have high economic value. The operation of these hybrid pumped-storage power stations requires the coordinated efforts of upstream and downstream reservoirs. When the pumped-storage units are pumping water for storage, sufficient water must be reserved in the downstream reservoirs. When the pumped-storage units are releasing water for power generation, sufficient storage capacity must be freed up in the downstream reservoirs. The reverse is true for the upstream reservoirs.

[0003] However, upstream and downstream stakeholders in my country's cascade hydropower stations often differ, leading to scheduling decisions by individual hydropower stations focused solely on maximizing their own interests. This lack of coordination between upstream and downstream stations can hinder the coordinated operation of multi-agent hybrid pumped-storage power plants, formed after capacity expansion and renovation of cascade hydropower stations with inconsistent upstream and downstream interests, thereby limiting the power generation efficiency of these plants. To ensure the efficient operation of multi-agent hybrid pumped-storage power plants, it is necessary to clarify the marginal contributions of different power generation entities to the system and calculate the compensation relationships between these entities.

[0004] Current research has largely focused on the sharing of benefits among multi-stakeholder hydropower stations, while insufficient attention has been paid to the distribution of benefits among the various generating entities within multi-stakeholder hybrid pumped-storage power stations. To meet the future needs of hybrid pumped-storage power stations for peak-shaving within the power grid, the coordinated operation and benefit distribution of hybrid pumped-storage power stations, formed through the integration and transformation of conventional hydropower stations, urgently require research. Summary of the Invention

[0005] The benefit distribution method under the joint scheduling of a multi-agent hybrid pumped-storage power station of the present invention is based on the constructed long-term and short-term nested scheduling model to fairly distribute the increased benefits of the coordinated operation of various power generation entities within the hybrid pumped-storage power station, thereby safeguarding the normal operation of the hybrid pumped-storage power station.

[0006] The technical solution adopted by the present invention is a profit distribution method under the joint dispatch of multi-agent hybrid pumped storage power stations, comprising the following steps:

[0007] Step 1: Obtain grid system load data to generate peak and valley electricity prices and daily average electricity prices;

[0008] Step 2: Establish and solve the medium- and long-term scheduling model of the hybrid pumped storage power station;

[0009] A medium- and long-term scheduling model for a hybrid pumped storage power station is established with the goal of maximizing peak-shaving power generation benefits, and a dynamic programming algorithm is used to solve it.

[0010] The specific formula for maximizing the peak-shaving power generation benefit is:

[0011] (6)

[0012] (7)

[0013] Where: For the i Hydropower stations are under medium- and long-term dispatch t Power generation during the time period; for t Average electricity price during the time period; T To calculate the total number of time periods; For the i Hydropower stations in t Output during the time period;

[0014] Step 3: Establish and solve the short-term dispatch model of the hybrid pumped storage power station;

[0015] Using the results of the medium- and long-term scheduling model as daily scheduling boundary conditions, the dynamic decision domain reduction technique is used to dynamically reduce the discrete space of reservoir water levels. Based on the SOS2 constraint, a short-term scheduling model for a hybrid pumped-storage power station is constructed with the goal of maximizing the intraday peak-shaving power generation efficiency. The short-term scheduling model is solved using the GUROBI linear solver.

[0016] The benefit of peak-shaving power generation during the day is the greatest, and the specific formula is as follows:

[0017] (twenty two)

[0018] Where: The first hybrid pumped storage power station i Hydropower station in Japan t Hourly power generation; For the pumped storage unit t The amount of electricity generated by pumping in hours; when “+” is used, it is hydropower generation; when “-” is used, it is pumping energy consumption; For the t Hourly peak and valley electricity prices;

[0019] Step 4: Allocate the benefits of the multi-agent hybrid pumped storage power station joint dispatch based on the Shapley value method;

[0020] Based on the constructed medium- and long-term scheduling model and short-term scheduling model of the hybrid pumped-storage power station, the power generation benefits of each power generation entity under different cooperative scenarios of individual scheduling, joint scheduling without pumped storage, and joint scheduling with pumped storage are calculated respectively. The Shapley value method based on cooperative game theory is used to distribute the benefits of each power generation entity in the hybrid storage system.

[0021] The present invention is also characterized in that:

[0022] The steps for generating peak and valley electricity prices and daily average electricity prices in step 1 are as follows:

[0023] Step 1.1: Obtain the daily load data of the power grid system in the previous year and divide the daily load data into data corresponding to high, medium and low periods;

[0024] The daily load data is divided into data corresponding to high, flat and valley periods. The method used is the three-equal division method. The three-equal division formula is as follows:

[0025] (1)

[0026] (2)

[0027] (3)

[0028] Where: and are the maximum and minimum loads of the day, respectively; The peak-to-valley difference of load in different time periods; 、 The critical points are divided into peak, flat and valley loads respectively.

[0029] Step 1.2: Set the base electricity price and the peak-valley electricity price ratio, and generate the corresponding peak-valley electricity price based on the daily load data division in step 1.1;

[0030] Step 1.3: Sort the peak and valley electricity prices corresponding to each hour of each day from high to low, calculate the average electricity price under different durations, and form the electricity price-duration curve. K The corresponding average electricity price is calculated as follows:

[0031] (4)

[0032] Where: For the duration K The corresponding average electricity price; It is the peak-valley electricity price sequence after rearrangement from high to low.

[0033] The constraints of the medium- and long-term scheduling model in step 2 are as follows:

[0034] a. Water balance:

[0035] (8)

[0036] Where: 、 Respectively i Hydropower stations in 、 t Reservoir water storage at the beginning of the time period; 、 Respectively i Hydropower stations in t Inbound and outbound traffic during the time period;

[0037] b. Downflow flow requirements:

[0038] (9)

[0039] Where: For the i The discharge flow restrictions required by each hydropower station during special periods;

[0040] c. Water level constraints:

[0041] (10)

[0042] Where: For the i Hydropower stations in t Reservoir water level during the period; 、 Respectively i Minimum and maximum water level constraints for each hydropower station;

[0043] d. Water level constraints at the beginning and end of the year:

[0044] (11)

[0045] (12)

[0046] Where: and Respectively i Reservoir water levels of each hydropower station at the beginning and end of the year;

[0047] e. Storage capacity constraints:

[0048] (13)

[0049] (14)

[0050] Where: 、 Respectivelyi Minimum and maximum storage capacity limits for each hydropower station; For the i Hydropower stations in t upstream water level during the time period; is the functional relationship between the corresponding dam front water level and reservoir capacity;

[0051] f. Traffic Constraints:

[0052] (15)

[0053] (16)

[0054] (17)

[0055] (18)

[0056] Where: 、 Respectively i Hydropower stations in t Power generation flow and water abandonment flow during the period; 、 Respectively i The minimum and maximum outflow rates of each hydropower station; 、 Respectively i The minimum and maximum power generation flow of each hydropower station; For the i Hydropower stations in t Downstream tailwater level during the time period; is the functional relationship between the discharge flow of the power station and the tailwater level;

[0057] g. Electric energy conversion relationship:

[0058] (19)

[0059] (20)

[0060] Where: Indicates the i The power generation function of a hydropower station, the power station output and the power flow and power generation water purification head Both are related; head loss Take it as a fixed value of 1m;

[0061] h. Output constraints:

[0062] (twenty one)

[0063] Where: For the i The forced output of a hydropower station is equal to the output of the station when it discharges at the ecological flow rate; For the i The maximum output of a hydropower station, that is, the installed capacity of the station.

[0064] In addition to the constraints mentioned in the medium- and long-term scheduling model (Equations (9), (10), (13) to (21)), the short-term scheduling model in step 3 also needs to meet the following additional constraints:

[0065] j. Water balance:

[0066] (twenty three)

[0067] Where: 、 Respectively i Hydropower stations in and t Reservoir water storage at the beginning of the time period; 、 Respectively i Hydropower stations in t Inbound and outbound traffic during the time period; and Respectively i Pumped storage units in t Pumping and power generation flows during the period;

[0068] k. Constraints on the water output of pumped storage units:

[0069] The additional pumped storage units are subject to the daily pumping water balance constraint:

[0070] (twenty four)

[0071] 1. Daily water level constraints:

[0072] (25)

[0073] (26)

[0074] Where: and Respectively i The reservoir water levels of each hydropower station at the beginning and end of the day are obtained from the medium- and long-term scheduling results;

[0075] m. Flow Constraints:

[0076] (27)

[0077] Where: 、 Respectively i The minimum and maximum pumping and discharging flow rates of each pumped storage unit;

[0078] n. Electric energy conversion relationship:

[0079] (28)

[0080] (29)

[0081] Where: For the i Pumped storage units in t Pumping power during the period; Indicates the i The pumping power conversion relationship of each pumped storage unit and the pumping flow rate and pumping head related; For the i Hydropower stations in t The head loss during the period; the head loss is also taken as a fixed value of 1m;

[0082] o. Output constraints:

[0083] (30)

[0084] Where: and Respectively i Pumped storage units in t Pumping and power generation capacity during the period; For the i The forced output of each pumped storage unit is set to 0; For the i The maximum pumped hydropower generation power of a pumped storage unit is the installed capacity.

[0085] Step 3 The calculation formula for the fluctuation range of reservoir water level is as follows:

[0086] (39)

[0087] (40)

[0088] Where: and Respectively i The amount of water stored in the reservoir corresponding to the highest and lowest water levels of each hydropower station during the day; For the i The reservoir capacity of each hydropower station corresponding to the water level at the beginning of the day; For the i The inflow of each hydropower station; and Respectively i The maximum pumping and power generation flow of pumped storage unit i.

[0089] The beneficial effects of the present invention are:

[0090] 1. The dynamic peak-valley electricity price proposed in this application changes in real time with system load, effectively guiding hybrid pumped-storage power plants to track and regulate load within a day. By allocating peak-shaving power to the power plant over long timescales through the power price-duration equation, this ensures sufficient peak-shaving power for the power plant's short-term operation and prevents the impact of medium- and long-term water allocation on the power plant's daily peak-shaving operation. Therefore, the dynamic peak-valley electricity price method proposed in this paper can guide hybrid pumped-storage power plants to perform peak-shaving operations on both long and short timescales.

[0091] 2. The long-term and short-term nested coupling scheduling model proposed in this paper with the maximization of power generation efficiency as the optimization goal can truly reflect the peak-shaving process of the multi-agent hybrid pumped-storage power station in the system; the short-term operation model is linearized by the SOS2 method, ensuring the solution accuracy of the complex scheduling model of the power station; in order to solve the connection problem of the long-term and short-term scheduling models, the dynamic decision domain reduction technology proposed in this paper can effectively reduce the solution space of the short-term model, thereby greatly reducing the solution time of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0092] Figure 1 It is a flow chart of the present invention;

[0093] Figure 2 is the result of generating the dynamic electricity price of the system of the present invention (typical day);

[0094] Figure 3 This is the dynamic electricity price generation process of the system of the present invention (all year round);

[0095] Figure 3 (a) is the net load distribution diagram of the present invention within the year;

[0096] Figure 3 (b) is the daily dynamic peak and valley electricity price distribution diagram of the present invention;

[0097] Figure 3 (c) is the electricity price-time distribution diagram of the present invention;

[0098] Figure 4 The discretization of the hydropower generation function of the present invention and the distribution of weight coefficients of each discrete point;

[0099] Figure 5 It is a schematic diagram of the dynamic decision domain reduction process of the present invention;

[0100] Figure 6This is the long-time-scale scheduling result of the hybrid pumped-storage power station of the present invention;

[0101] Figure 7 This is the operation status of the hybrid pumped storage power station of the present invention on a typical day during the flood season;

[0102] Figure 8 This is the operation status of the hybrid pumped storage power station of the present invention on a typical day during the dry season;

[0103] Figure 9 This is the daily change process of the water level before and after the pumped storage operation of the A hydropower station in the present invention;

[0104] Figure 10 This is the daily change process of the water level before and after the pumped storage operation of the B hydropower station in the present invention;

[0105] Figure 11 This is the change in power generation efficiency of hydropower station A and hydropower station B under different cooperation modes in the present invention. DETAILED DESCRIPTION

[0106] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0107] The benefit distribution method under the joint dispatch of multi-agent hybrid pumped storage power station of the present invention, Figure 1 is a flow chart of the present invention;

[0108] The mixed storage power station has hydropower station A upstream and hydropower station B downstream, and at least one pumped storage unit is set between hydropower station A and hydropower station B.

[0109] The specific steps are as follows:

[0110] Step 1: Obtain the load data of the power grid system and generate the peak and valley electricity prices and the daily average electricity price;

[0111] Unlike the traditional goal of maximizing power generation, in the new power system dominated by new energy, hybrid pumped-storage power stations mainly play the role of peak regulation and absorbing new energy. The operation of the power station must adapt to the changes in the system's net load. At this time, the use of traditional fixed electricity prices can no longer fully utilize the flexible adjustment capabilities of the power station.

[0112] To this end, this invention generates dynamic peak-valley electricity prices based on the net load fluctuations of the Shaanxi power grid in a given year, after deducting renewable energy from the actual load. This process simultaneously guides hybrid pumped-storage power plants to regulate the system's peak load on both long and short time scales. The generation of dynamic peak-valley electricity prices involves three steps: generating an intraday peak-valley price curve, a price-over-time curve, and a daily average peak-shaving price.

[0113] (1) Intraday peak and valley electricity price curve

[0114] First, the load is divided into three equal parts to divide the daily load into peak period, valley period, and mid-load period, which is the normal period. The load division process is shown in equations (1) to (3):

[0115] (1)

[0116] (2)

[0117] (3)

[0118] Where: and Corresponding to the maximum and minimum net load of the day, 10 4 kW; is the load peak-valley difference in different time periods, 10 4 kW; 、 They are the critical points for dividing the load peak, valley and mid-load periods, 10 4 kW.

[0119] By setting the above three types of loads to correspond to one electricity price level respectively, the daily electricity price process can be divided into three stages: peak, flat and valley that match the load process.

[0120] (2) Electricity price history curve

[0121] Arrange the obtained daily peak and valley electricity prices from high to low, calculate the average electricity price under different durations, and form the electricity price-duration curve. K The corresponding average electricity price is calculated as follows:

[0122] (4)

[0123] Where: For the duration K The corresponding average electricity price is RMB / kW·h; It is the peak-valley electricity price sequence after rearrangement from high to low.

[0124] (3) Daily average peak-shaving electricity price

[0125] First, calculate the daily installed capacity of the power station The maximum adjustable time during peak load operation , the calculation formula is as follows:

[0126] (5)

[0127] Where: is the peak-shaving power of the power station during the daytime peak-shaving operation, 10 4 kW·h; is the daily power generation of the power station when it is operating at forced output, 10 4 kW·h; is the installed capacity of the power station unit, 10 4 kW; is the forced output of the power station. This article shows the output of the power station when it discharges at the ecological flow rate. 4 kW.

[0128] According to the maximum peak load regulation time of the power station t By checking the daily electricity price-duration curve obtained above, we can get the average electricity price of the power station during peak load operation on that day. .

[0129] If the base electricity price is set at 0.3 yuan / kW·h and the peak-to-valley ratio of electricity price is 3, the daily peak electricity price will be 0.45 yuan / kW·h and the valley electricity price will be 0.15 yuan / kW·h. Figure 2 The peak-valley electricity price and electricity price-duration curve generated based on the system net load process on a typical day are shown. Figure 2 The peak-valley electricity price on this typical day is consistent with the distribution law of the system net load, showing a bimodal distribution feature; the average electricity price gradually decreases from the peak electricity price of 0.45 yuan / kW·h to the daily average electricity price of 0.31 yuan / kW·h as the electricity price increases over time. After lasting for more than 10 hours, the unit peak-shaving benefit of the power station gradually weakens.

[0130] According to the net load distribution of the system during the year Figure 3 (a) in the above equation is generated as Figure 3 Based on the dynamic peak-valley electricity price distribution map of the year, the above method is used to create the electricity price-duration curve, which is generated as follows: Figure 3 (c) The annual electricity price-time distribution diagram. The results show that Figure 3 In (a), the distribution of the system net load at different times of the year shows its own unique pattern. Therefore, the peak-valley electricity price must be flexible and changeable when guiding hydropower stations and other power plants to carry out peak regulation. Figure 3 In (b), the valley electricity prices in each period are basically between 0:00 and 6:00, while the peak electricity prices show different distribution characteristics throughout the year; Figure 3 The results in (c) show that the system has a higher demand for peak load in summer and winter, but less demand in spring and autumn, which is consistent with the seasonal distribution characteristics of the load.

[0131] The results show that through the constructed dynamic peak-valley electricity price generation model, not only can the duration corresponding to the daily peak-shaving benefit threshold be obtained, but the power station can retain the excess water to be discharged on other days to obtain a higher peak-shaving electricity price; it can also obtain the distribution characteristics of net load, peak-valley electricity prices and electricity demand in different seasons of the year, thereby achieving the balance of peak-shaving electricity on a long-term scale.

[0132] Step 2: Establish and solve the medium- and long-term scheduling model of the hybrid pumped storage power station;

[0133] In order to minimize the impact of the operation of pumped-storage units on the original hydropower station, the multi-subject hybrid pumped-storage power station sets the pumped-storage units to balance the water pumping within the day. In other words, the medium- and long-term scheduling process of the power station is actually just the optimized scheduling for conventional cascade hydropower stations.

[0134] 1. Objective Function

[0135] Taking into account the peak-shaving operation of the power station, the medium- and long-term power generation benefits of the hybrid pumped-storage power station are calculated by multiplying the daily power generation of the cascade hydropower station by the average electricity price of the day:

[0136] (6)

[0137] (7)

[0138] Where: For the i Hydropower stations are under medium- and long-term dispatch t Power generation during the period, 10 4 kW·h; For the t Average electricity price during the time period, yuan / kW·h; T To calculate the total number of time periods; For the i Hydropower stations in t Output of the period, 10 4 kW; medium and long-term scheduling calculation period Take as day, and the total calculation duration as year. I represents the total number of hydropower stations.

[0139] 2. Constraints

[0140] a. Water balance:

[0141] (8)

[0142] Where: 、 Respectively i Hydropower stations in 、 t The reservoir water storage at the beginning of the period, 10 8 m3 ; 、 Respectively i Hydropower stations in t Inbound and outbound traffic during the period, m 3 / s.

[0143] b. Downflow flow requirements:

[0144] (9)

[0145] Where: For the i The discharge flow limit required by each hydropower station during special periods, m 3 / s.

[0146] c. Water level constraints:

[0147] (10)

[0148] Where: For the i Hydropower stations in t Reservoir water level during the period, m; 、 Respectively i The minimum and maximum water level constraints of each hydropower station, m.

[0149] d. Water level constraints at the beginning and end of the year:

[0150] (11)

[0151] (12)

[0152] Where: and Respectively i Reservoir water levels of hydropower stations at the beginning and end of the year, m.

[0153] e. Storage capacity constraints:

[0154] (13)

[0155] (14)

[0156] Where: 、 Respectively i The minimum and maximum storage capacity limits of hydropower stations, 10 8 m 3 ; For the i Hydropower stations in tupstream water level during the period, m; is the corresponding functional relationship between the water level in front of the dam and the reservoir capacity.

[0157] f. Traffic Constraints:

[0158] (15)

[0159] (16)

[0160] (17)

[0161] (18)

[0162] Where: 、 Respectively i Hydropower stations in t Power generation flow and abandoned water flow in the period, m 3 / s; 、 Respectively i The minimum and maximum outflow of each hydropower station, m 3 / s; 、 Respectively i The minimum and maximum power generation flow of each hydropower station, m 3 / s; For the i Hydropower stations in t Downstream tailwater level of the time period, m; It is the functional relationship between the discharge flow of the power station and the tailwater level.

[0163] g. Electric energy conversion relationship:

[0164] (19)

[0165] (20)

[0166] Where: Indicates hydropower station i The power generation function, power station output and power flow and power generation water purification head Both are related; head loss Take it as a fixed value of 1m.

[0167] h. Output constraints:

[0168] (twenty one)

[0169] Where: For the iThe forced output of a hydropower station is equal to the output of the station when it discharges at the ecological flow rate, 10 4 kW; For the i The maximum output of a hydropower station, i.e. the installed capacity of the station, 10 4 kW.

[0170] In the constraints of the above medium- and long-term scheduling model, the calculation period t is taken as day.

[0171] The solution to the medium- and long-term scheduling model adopts the conventional dynamic programming algorithm, which will not be described here.

[0172] Step 3: Establish and solve the short-term dispatch model of the hybrid pumped storage power station;

[0173] The medium- and long-term scheduling results of the hybrid pumped storage power station are input into the short-term scheduling model as hard constraints to achieve nested coupling of the medium- and long-term scheduling model and the short-term scheduling model.

[0174] 1. Objective Function

[0175] The short-term dispatch target is obtained by multiplying the peak and valley electricity prices within the day by the power plant's hourly power withdrawal:

[0176] (twenty two)

[0177] Where: The first hybrid pumped storage power station i Hydropower station in Japan t Hours of power generation, 10 4 kW·h; For the pumped storage unit t When “+” is taken, it is hydropower generation, and when “-” is taken, it is pumping energy consumption. 4 kW·h; For the t The peak-valley electricity price for the hour is RMB / kW·h. The power generation of the power station per hour at this time is and output size equal.

[0178] 2. Constraints

[0179] The calculation period of the short-term dispatch model of the hybrid pumped storage power station is taken as hours, and the total calculation time is taken as days. In addition to the constraints mentioned in the medium- and long-term dispatch model (Equations (9), (10), (13) to (21)), the following additional constraints need to be met:

[0180] j. Water balance:

[0181] Different from the medium- and long-term scheduling process, the pumping and power generation flows of the pumped storage units between upstream and downstream power stations need to be considered during short-term scheduling. The water balance calculation formula for the power station is as follows:

[0182] (twenty three)

[0183] Where: 、 Respectively i Hydropower stations in and t The reservoir water storage at the beginning of the period, 10 8 m 3 ; 、 Respectively i Hydropower stations in t Inbound and outbound traffic during the period, m 3 / s; and Respectively i Pumped storage units in t Pumping and power generation flow rate during the period, m 3 / s.

[0184] k. Constraints on the water output of pumped storage units:

[0185] The additional pumped storage units are subject to the daily water pumping balance constraints.

[0186] (twenty four)

[0187] 1. Daily water level constraints:

[0188] (25)

[0189] (26)

[0190] Where: and Respectively i The reservoir water levels of each hydropower station at the beginning and end of the day are obtained from the medium- and long-term scheduling results, m.

[0191] m. Flow Constraints:

[0192] (27)

[0193] Where: 、 Respectively i The minimum and maximum pumping and discharging flow of each pumped storage unit, m 3 / s.

[0194] n. Electric energy conversion relationship:

[0195] (28)

[0196] (29)

[0197] Where: For the i Pumped storage units in t Pumping power for the period, 10 4 kW; Indicates the i The pumping power conversion relationship of each pumped storage unit and the pumping flow rate and pumping head Related: Pumping head By upstream and downstream power stations t The difference in reservoir water levels during each time period plus the head loss is calculated as m. When the pumped storage unit discharges water for power generation, its power generation function is the same as that of the conventional hydropower units at Hydropower Station A and will not be further described here. The head loss is also taken as a fixed value of 1 m.

[0198] o. Output constraints:

[0199] (30)

[0200] Where: and Respectively i Pumped storage units in t Pumping and power generation during the period, 10 4 kW; For the i The forced output of the pumped storage unit is 0, 10 4 kW; For the i The maximum pumped hydropower generation capacity of each pumped storage unit, i.e. installed capacity, is 10 4 kW.

[0201] The scheduling process of cascade hydropower is itself a non-convex nonlinear optimization problem. The addition of pumped storage units further increases the complexity of the model solution. Considering that the scheduling objectives and optimization subjects in medium- and long-term scheduling and short-term scheduling are different, this paper adopts two methods to calculate them respectively. Among them, the solution of the medium- and long-term scheduling model adopts a conventional dynamic programming algorithm, which will not be repeated here. The short-term scheduling model uses the SOS2 method to linearize the nonlinear power generation function into a mixed integer problem for solution. In addition, in order to reduce the calculation time of the short-term model, the present invention also proposes a dynamic decision domain reduction technology for the mixed integer programming algorithm.

[0202] (1) Linearization of power generation function

[0203] The SOS2 constraint is a special ordered set constraint that represents a continuous variable set in which at most two adjacent variables are not 0 and the rest of the variables are 0. The steps to use it to express the hydropower generation function are as follows:

[0204] Step 1: Build Relationship surface is used to represent the relationship between the power generation of the power station and the power flow and the water level of the reservoir. V and traffic Q A discrete point sequence 、 , then the power generation P Available through Spatial distribution points Approximately, each All and only and Correspondingly, ,in M and N Represent the discrete totals of storage capacity and power generation flow respectively.

[0205] Step 2: V and Q Discrete point combination , Add SOS2 constraints to ensure that in any period, V and Q Can only pass through a single interval and Approximately expressed, and this interval also corresponds to the only power generation P .in, , .

[0206] Step 3: Create continuous variables using weight coefficients V 、 Q 、 P Same as discrete variables 、 、 The corresponding relationship. Introducing auxiliary variables 、 as well as Respectively represent storage capacity V , power generation flow Q , and power generation P The weights assigned to different vertices of the discrete mesh are as follows: Figure 4 shown.

[0207] but V 、 Q 、 P They can be expressed by the following formulas:

[0208] (31)

[0209] (32)

[0210] (33)

[0211] Among them, the power generation in each grid P The sum of the weights of each discrete point should be equal to 1, and x The sum of the weights of the discrete points in the axial direction is equal to the storage capacity The weight coefficient of y The sum of the weights of the discrete points in the axial direction is equal to the downstream flow rate The weight coefficient is:

[0212] (34)

[0213] (35)

[0214] (36)

[0215] To ensure continuous variables V 、 Q The uniqueness of the distribution interval, and The SOS2 constraints should be satisfied respectively, namely:

[0216] (37)

[0217] (38)

[0218] (2) Dynamic decision domain reduction technology

[0219] The short-term dispatch calculation period of a hybrid pumped storage power station is hourly, and the total calculation time is one year. The calculation scale is large and time-consuming. Considering that during short-term operation, the water level of the power station reservoir fluctuates around the medium- and long-term dispatch water level process line, this paper uses the short-term fluctuation range of the reservoir water level to calculate the discrete interval of the power station's daily operating water level corresponding to the storage capacity. Dynamic reduction is performed to eliminate invalid decision spaces and shorten calculation time.

[0220] The calculation formula for the daily fluctuation range of reservoir water level is as follows:

[0221] (39)

[0222] (40)

[0223] Where: and Respectively i The corresponding reservoir water storage capacity of the hydropower station at the highest and lowest water levels of the day, 10 8 m 3 ; For the i The reservoir capacity of each hydropower station corresponding to the water level at the beginning of the day, 10 8 m 3 ; For the i Inflow of each hydropower station, m 3 / s; and Respectively i The maximum pumping and power generation flow of pumped storage unit i, m 3 / s.

[0224] Similarly, the daily discharge flow interval of the power station can also be discretized according to the upper and lower limits of the discharge flow. After the decision domain is reduced, the optimization space of the short-term scheduling model can be greatly reduced, and the calculation time can be reduced. The principle of dynamic decision domain reduction is shown in Figure 5 shown.

[0225] like Figure 6 As shown in the results, the proposed maximum peak-shaving benefit model can reduce the discharge of Hydropower Station A during periods of high water inflow, such as July and August, while increasing the discharge during periods of low water inflow, such as March and April, compared to the maximum power generation model. This alleviates the uneven discharge process of Hydropower Stations A and B throughout the year and ensures sufficient peak-shaving water volume for the hybrid pumped-storage power station during different seasons. As shown in Table 1, when Hydropower Station A uses the maximum peak-shaving benefit model for medium- and long-term scheduling, its annual power generation is 2.63% lower and its total power generation benefit is 3.32% higher than when using the maximum power generation model. The results show that the dynamic peak-valley electricity price proposed in this paper can effectively improve the peak-shaving capacity of the hybrid pumped-storage power station.

[0226] Table 1A Hydropower Station Power Generation (Compared with Maximum Power Generation Dispatching Target)

[0227]

[0228] like Figure 7 and Figure 8 The typical daily operation process of the hybrid pumped storage power station in the flood season and the dry season is shown in Figure 2. It can be found that the output of conventional hydropower on a typical day in the flood season is much greater than that on a typical day in the dry season. However, the operation mode of the pumped storage unit at maximum power does not change under the two water flow rates. Figure 9 and Figure 10The daily changes in reservoir water levels at Hydropower Stations A and B before and after pumped-storage operation are shown. It's clear that the daily water level at Hydropower Station A has generally risen, up to 0.25 meters compared to the pre-pumped-storage period, due to the pumping and power generation processes of the pumped-storage units. Meanwhile, the daily water level at Hydropower Station B has generally declined, with a maximum drop of 0.75 meters. Statistics show that after the pumped-storage units began operating, the average annual generated hydraulic head at Hydropower Station A increased by 0.14% compared to before the pumped-storage units began operating, while that at Hydropower Station B decreased by 0.20%. This change may affect the power generation efficiency of Hydropower Stations A and B.

[0229] Step 4: Allocate the benefits of the multi-agent hybrid pumped storage power station joint dispatch based on the Shapley value method;

[0230] The Shapley value method is a cooperative solution that considers the weighted average of all participants' contributions to each alliance as the fair and effective benefit of the participant under the cooperative mode. There are n participants in the CCP. Considering that different participants join the alliance in different orders, we set For all permutation order combinations, then The combination may have When the participant i is in a permutation order After joining the alliance, use represents the set of entities that joined the alliance before participant i. The marginal contribution is calculated as follows:

[0231] (41)

[0232] The formula for calculating the Shapley value of participant i in the cooperative game is as follows:

[0233] (42)

[0234] The benefit that participant i should share in the cooperative alliance is:

[0235] (43)

[0236] Where: N is the alliance set composed of n participants in the cooperative game; S is the set of entities that join the alliance before participant i in a permutation order; is the number of members in the set; is the total benefit of the collection; For collection S The total benefit after adding participant i; Indicates that before subject i joins the alliance, The number of permutations and combinations of members entering the alliance, After subject i joins the alliance, the remaining The number of permutations and combinations of members entering the alliance.

[0237] Based on the constructed medium- and long-term nested dispatch model, the power generation benefits of three different power generation entities, hydropower station A, hydropower station B, and pumped storage units, are calculated under six cooperative scenarios: independent dispatch, joint dispatch without pumped storage, and joint dispatch with pumped storage, as shown in Table 2:

[0238]

[0239] On this basis, the Shapley value method based on cooperative game theory is used to calculate the marginal contribution of each power generation entity in the hybrid storage system to the system joint gain. Finally, the compensation relationship between each power generation entity is determined, as shown in Table 3. Positive values ​​are compensated, and negative values ​​are compensation. The unit is 100 million yuan.

[0240]

[0241] According to the compensation relationship in Table 3, before the pumped-storage unit was operational, Hydropower Station A increased Hydropower Station B's power generation efficiency by approximately 18% by regulating inflow, while losing 0.1% of its own efficiency. Therefore, before the pumped-storage unit was constructed, the downstream Hydropower Station B was obligated to compensate the upstream Hydropower Station A. After the pumped-storage unit was operational, the overall system's power generation efficiency increased from 1.2923 billion yuan without the pumped-storage unit to 1.4686 billion yuan, while Hydropower Station A's revenue decreased even more than before the pumped-storage unit was constructed. Using the Shapley value method to allocate the increased revenue from the hybrid pumped-storage power station's integration, the pumped-storage unit is required to compensate the upstream Hydropower Station A for 91 million yuan and the downstream Hydropower Station B for 27.9 million yuan.

[0242] from Figure 11 As can be seen, after the pumped-storage system was put into operation, the revenue levels of both Hydropower Station A and Hydropower Station B were higher than before the system was put into operation, and both were higher than when the two were operated independently. While the pumped-storage units lost a significant portion of their revenue, they ensured that upstream and downstream power stations proactively adjusted their operating methods to maintain normal operation. Therefore, this benefit distribution result can achieve a win-win situation for the different power generation entities within the hybrid pumped-storage power station.

Claims

1. A method for distributing benefits under the joint dispatch of a multi-agent hybrid pumped storage power station, characterized in that: The following steps are involved: Step 1: Obtain grid system load data to generate peak and valley electricity prices and daily average electricity prices; The steps for generating peak and valley electricity prices and daily average electricity prices are as follows: Step 1.1: Obtain the daily load data of the power grid system in the previous year and divide the daily load data into data corresponding to high, medium and low periods; Step 1.2: Set the base electricity price and the peak-valley electricity price ratio, and generate the corresponding peak-valley electricity price based on the daily load data division in step 1.1; Step 1.3: Sort the peak and valley electricity prices corresponding to each hour of each day from high to low, calculate the average electricity price under different durations, and form an electricity price-duration curve; where duration K The corresponding average electricity price is calculated as follows: (4) Where: For the duration K The corresponding average electricity price; is the peak-valley electricity price sequence rearranged from high to low; Calculate the daily installed capacity of the power station The maximum adjustable time during peak load operation , the calculation formula is as follows: (5) Where: is the peak-shaving power of the power station during the daytime peak-shaving operation, 10 4 kW·h; is the daily power generation of the power station when it is operating at forced output, 10 4 kW·h; is the installed capacity of the power station unit, 10 4 kW; For the forced output of the power station, 10 4 kW; According to the maximum peak load regulation time of the power station t By checking the daily electricity price-duration curve obtained above, we can get the average electricity price of the power station during peak load operation on that day. ; Step 2: Establish and solve the medium- and long-term scheduling model of the hybrid pumped storage power station; A medium- and long-term scheduling model for a hybrid pumped storage power station is established with the goal of maximizing peak-shaving power generation benefits, and a dynamic programming algorithm is used to solve it. The specific formula for maximizing the peak-shaving power generation benefit is: (6) (7) Where: For the i Hydropower stations are under medium- and long-term dispatch t Power generation during the time period; for t Average electricity price during the time period; T To calculate the total number of time periods; For the i Hydropower stations in t The output of the period; I is the total number of hydropower stations, For medium to long term calculation period; Step 3: Establish and solve the short-term dispatch model of the hybrid pumped storage power station; Using the results of the medium- and long-term scheduling model as daily scheduling boundary conditions, the dynamic decision domain reduction technology is used to dynamically reduce the discrete space of the reservoir water level. Based on the SOS2 constraint, a short-term scheduling model for a hybrid pumped-storage power station is constructed with the goal of maximizing the intraday peak-shaving power generation benefit. The short-term scheduling model is solved using the GUROBI linear solver. The intraday peak-shaving power generation benefit is the greatest, and the specific formula is as follows: (22) Where: The first hybrid pumped storage power station i Hydropower station in Japan t Hourly power generation; For the pumped storage unit t The amount of electricity generated by pumping per hour; when "+" is taken, it is hydropower generation, and when "-" is taken, it is pumping energy consumption; For the t Hourly peak and valley electricity prices; The calculation formula for the fluctuation range of reservoir water level is as follows: （39） （40） Where: and Respectively i The amount of water stored in the reservoir corresponding to the highest and lowest water levels of each hydropower station during the day; For the i The reservoir capacity of each hydropower station corresponding to the water level at the beginning of the day; For the i The inflow of each hydropower station; and Respectively i The maximum pumping and power generation flow of each pumped storage unit i; Step 4: Allocate the benefits of the multi-agent hybrid pumped storage power station joint dispatch based on the Shapley value method; Based on the constructed medium- and long-term scheduling model and short-term scheduling model of the hybrid pumped-storage power station, the power generation benefits of each power generation entity under different cooperative scenarios of individual scheduling, joint scheduling without pumped storage, and joint scheduling with pumped storage are calculated respectively. The Shapley value method based on cooperative game theory is used to distribute the benefits of each power generation entity in the hybrid storage system.

2. The profit distribution method under the joint dispatch of multi-agent hybrid pumped storage power station according to claim 1 is characterized in that: In step 1.1, the daily load data is divided into data corresponding to high, medium and low periods. The method used is the three-equal division method. The three-equal division formula is as follows: （1） （2） （3） Where: and are the maximum and minimum loads of the day, respectively; The peak-to-valley difference of load in different time periods; 、 The critical points are divided into peak, flat and valley loads respectively.

3. The profit distribution method under the joint dispatch of multi-agent hybrid pumped storage power station according to claim 1 is characterized in that: The constraints of the medium- and long-term scheduling model described in step 2 are as follows: a. Water balance: (8) Where: 、 Respectively i Hydropower stations in 、 t Reservoir water storage at the beginning of the time period; 、 Respectively i Hydropower stations in t Inbound and outbound traffic during the time period; b. Downflow flow requirements: (9) Where: For the i The discharge flow restrictions required by each hydropower station during special periods; c. Water level constraints: (10) Where: For the i Hydropower stations in t Reservoir water level during the period; 、 Respectively i Minimum and maximum water level constraints for each hydropower station; d. Water level constraints at the beginning and end of the year: (11) (12) Where: and Respectively i Reservoir water levels of each hydropower station at the beginning and end of the year; e. Storage capacity constraints: (13) (14) Where: 、 Respectively i Minimum and maximum storage capacity limits for each hydropower station; For the i Hydropower stations in t upstream water level during the time period; is the functional relationship between the corresponding dam front water level and reservoir capacity; f. Traffic Constraints: (15) (16) (17) (18) Where: 、 Respectively i Hydropower stations in t Power generation flow and water abandonment flow during the period; 、 Respectively i The minimum and maximum outflow rates of each hydropower station; 、 Respectively i The minimum and maximum power generation flow of each hydropower station; For the i Hydropower stations in t Downstream tailwater level during the time period; is the functional relationship between the discharge flow of the power station and the tailwater level; g. Electric energy conversion relationship: (19) (20) Where: Indicates the i The power generation function of a hydropower station, the power station output and the power flow and power generation water purification head Both are related; head loss Take it as a fixed value of 1m; h. Output constraints: (21) Where: For the i The forced output of a hydropower station is equal to the output of the station when it discharges at the ecological flow rate; For the i The maximum output of a hydropower station, that is, the installed capacity of the station.

4. The profit distribution method under the joint dispatch of multi-agent hybrid pumped storage power station according to claim 3 is characterized in that: In addition to the constraints mentioned in the medium- and long-term scheduling models, the short-term scheduling model described in step 3 also needs to meet the following additional constraints: j. Water balance: (23) Where: 、 Respectively i Hydropower stations in and t Reservoir water storage at the beginning of the time period; 、 Respectively i Hydropower stations in t Inbound and outbound traffic during the time period; and Respectively i Pumped storage units in t Pumping and power generation flows during the period; k. Constraints on the water output of pumped storage units: The additional pumped storage units are subject to the daily pumping water balance constraint: (24) 1. Daily water level constraints: (25) (26) Where: and Respectively i The reservoir water levels of each hydropower station at the beginning and end of the day are obtained from the medium- and long-term scheduling results; m. Flow Constraints: (27) Where: 、 Respectively i The minimum and maximum pumping and discharging flow rates of each pumped storage unit; n. Electric energy conversion relationship: (28) (29) Where: For the i Pumped storage units in t Pumping power during the period; Indicates the i The pumping power conversion relationship of each pumped storage unit and the pumping flow rate and pumping head related; For the i Hydropower stations in t The head loss during the period; the head loss is also taken as a fixed value of 1m; o. Output constraints: (30) Where: and Respectively i Pumped storage units in t Pumping and power generation capacity during the period; For the i The forced output of each pumped storage unit is set to 0; For the i The maximum pumped hydropower generation power of a pumped storage unit is the installed capacity.

Images