A risk analysis and control method for pre-flood energy storage of cascaded water-wind-solar hybrid system

Abstract

The present invention belongs to the field of power generation dispatching of electric power system, and discloses a risk analysis and control method for pre-flood energy storage of cascaded water-wind-solar complementary system. Taking pre-flood energy storage as a constraint, a dry season drawdown model and flood season water storage dispatching rules are established to determine the simulation operation criteria; a set of indicators for power shortage in dry season, water abandonment in flood season, insufficient energy storage at the end of the year, and wind-solar abandoned power are constructed to quantify multi-stage multi-source operation risks; Monte Carlo and fuzzy membership are coupled to characterize multi-dimensional uncertainty scenarios and their probabilities, and simulation analysis is used to determine the quantitative relationship between pre-flood energy storage and water-wind-solar complementary benefits, risk probability, and risk loss. Through the example analysis of a water-wind-solar complementary system in a particularly large river basin, the results show that the present invention can comprehensively describe risks in terms of occurrence probability and loss. The verification results show that through accurate risk description and reasonable energy storage control, the system power generation increased by 580 million kWh, and the risk loss was reduced by an average of 42%, showing good practicality.

Description

Technical Field

[0001] The present invention belongs to the field of power system dispatching and relates to a cascade pre-flood energy storage risk analysis and control method for a water-wind-solar complementary system. Background Art

[0002] Making full use of my country's 100 million kilowatt-level hydropower advantages to carry out water-wind-solar complementarity is a realistic and reliable choice for my country to alleviate the flexibility demand of wind-solar and promote the consumption of wind-solar. However, in the complementary operation, how to coordinate the energy storage arrangements of cascade hydropower stations, especially large-scale control-type power stations, will greatly affect the compensatory support role of hydropower for new energy power and electricity. There are mainly three methods for the long-term dispatch and operation of hydropower in traditional water-wind-solar systems: 1) Optimize the long-term energy storage arrangement of hydropower from the perspective of long-term complementary electricity fluctuations and system benefits; 2) Consider the short-term demand and short-term operation characteristics of new energy to improve the reliability of long-term plans; 3) Introduce risk indicators as optimization criteria to reduce long-term planning risks. At present, long-term plans mostly focus on single or a few risk sources in terms of risk control (Wang Jin, Zhao Zhipeng, Cheng Chuntian, et al. Research on cascade hydro-wind-photovoltaic complementary scheduling rules based on coupled output damage depth and power abandonment criteria [J]. Journal of Hydraulic Engineering, 2023, 54(12): 1415-1429.), while the actual operation of hydro-wind-photovoltaic complementary systems faces complex multi-source risks, such as power shortage in existing systems (Guo Y, Ming B, Huang Q, et al. Risk-averse day-ahead generation scheduling of hydro-wind-photovoltaic complementary systems considering the steady requirement of power delivery [J]. Applied Energy, 2022, 309: 118467.), new energy curtailment (Ming Bo, Li Yan, Liu Pan, et al. Study on medium- and long-term optimal dispatch of hydropower-solar complementary power generation with embedded short-term curtailment risk [J]. Journal of Hydraulic Engineering, 2021, 52(06): 712-722.), hydropower station curtailment (Cao Rui, Cheng Chuntian, Shen Jianjian, et al. Long-term power generation dispatch method of power station considering water curtailment risk during water storage period. Journal of Hydraulic Engineering, 2021. 52(10): 1193-1203.), unreasonable cascade energy storage control (Niu Wenjing, Wu Xinyu, Feng Zhongkai, et al. Optimal dispatch method of energy storage control for cascade hydropower station group. Proceedings of the CSEE, 2017. 37(11): 3139-3147+3369.), etc.; and the current methods rarely consider the risk balance control needs in actual production operation. Therefore, it is particularly necessary to further quantify the multiple risks of hydropower-wind-solar complementary operation and propose risk control methods that are suitable for them.

[0003] Pre-flood energy storage control of cascade power stations is a key task for long-term scheduling, which directly affects the available water volume during the dry season and water storage during the flood season. The whole year can be divided into two main stages with the pre-flood season as the time node: the drawdown period (approximately January-June) and the water storage adjustment period (approximately July-December). Due to the seasonal characteristics of runoff and wind and solar energy, the operating characteristics of different stages put forward conflicting control requirements for cascade pre-flood energy storage control. Irrational control is very likely to cause key risks such as power shortage during the dry season, water abandonment during the flood season, wind and solar power abandonment, and unreasonable energy storage control at the end of the year; therefore, accurately grasping the relationship between pre-flood energy storage and various operational risks is conducive to the balanced control of multiple risks of complementary systems.

[0004] In view of the above problems, the present invention proposes a risk analysis and control method for pre-flood energy storage in a cascade of water-wind-solar complementary systems, and conducts an application test on the water-wind-solar complementary system in a certain large river basin as an engineering background. The results show that the results of the present invention can effectively quantify the fine relationship between the flood season energy storage of the cascade of the complementary system and various risks, and can effectively reduce system risks and reduce the losses of water-wind-solar complementary operation in actual application. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a risk analysis and control method for cascade pre-flood energy storage in a water-wind-solar complementary system, so as to characterize the various risks and losses faced by pre-flood energy storage, quantitatively analyze the multi-dimensional operation risks of the water-wind-solar complementary system, and reasonably control the cascade pre-flood energy storage.

[0006] Technical solution of the present invention:

[0007] A method for analyzing and controlling the risk of cascade energy storage before flood season in a water-wind-solar complementary system comprises the following steps:

[0008] (1) Taking the cascade pre-flood energy storage, monthly power control during the dry season, and conventional restrictions on hydropower dispatch as constraints, a dry season drawdown optimization model is constructed to obtain the optimal dry season drawdown scheme. The objective function of the dry season drawdown optimization model is as follows:

[0009]

[0010] Where E is the expected power generation under each scenario; J, T1, N represent the runoff uncertainty scenario, the number of dry season scheduling periods and the set of power stations, respectively, and j, t, n are the corresponding set elements; P j is the jth scene probability; Ph j,n,t is the output of power station n in the tth period of scenario j; Pwp j,t is the wind and solar power output in the tth period of scene j; Δt is the corresponding duration of the period (s);

[0011] The energy storage constraints before the flood season and the monthly power control constraints during the dry season are as follows:

[0012] 1) Constraints on cascade energy storage before flood season

[0013]

[0014] Where: To store energy before the flood season for power station n; E tar To store energy for the target before the flood season; Refers to the storage capacity of power station n at the end of the period; Refers to the total storage capacity of the power stations upstream of power station n, is the set of upstream power stations of power station n; η n is the average water consumption rate of power station n;

[0015] 2) Electricity control constraints during dry season months

[0016]

[0017] Where: K t is the control ratio of power generation in period t to dry period (historical statistics); ξ is the control error;

[0018] Using Gurobi solver as the modeling and solving platform, Python programming and pyomo modeling language are used to linearize the nonlinear constraints in the dry season drawdown optimization model and convert it into mixed integer linear programming for solving;

[0019] (2) Construct water storage scheduling rules, including five-stage hedging scheduling rules and cascade output allocation method, and the cascade output allocation method adopts K value discrimination method;

[0020] (2.1) Establish hedging dispatching rules: Use Python-pyomo modeling program to establish parameter linear optimization method to construct five-stage hedging dispatching rules to determine the monthly output of the water-wind-solar complementary system. The principle is shown in Figure 1 . In the figure, OAGD is the three-stage dispatching rule of the water-wind-solar complementary system, in which OA is the output section below the guaranteed output; AG is the output guarantee section; and GD is the increased output section. The hedging dispatching rule of the present invention is based on the three-stage dispatching rule, and adds points B and C between the AG and GD sections respectively, to establish the hedging dispatching rule OABCD containing the BC hedging section, in which the horizontal coordinates of the intersection of the BC section with AG and GD are parameters a and b. The method for determining parameters a and b is optimized based on the historical output results of water, wind and solar power, and the specific expressions are as follows:

[0021]

[0022] Where: k and d are the slope and intercept of the BC segment curve; E t is the average available energy of historical hydropower and wind power in period t; T2 is the number of scheduling periods during the water storage adjustment period; P tis the historical average output of the corresponding period t; since BC intersects with the AB segment and the GD segment, the expressions of the AB segment and the GD segment are combined, and a and b are calculated by the following formula:

[0023]

[0024] Where: P M It is the maximum output of the water-wind-solar complementary system; P f It is the guaranteed output of the water-wind-solar complementary system; is the fully available storage capacity of the cascade power station (excluding dead storage capacity); Δm is the number of hours in a month;

[0025] (2.2) Cascade output allocation: After obtaining the monthly output of the hydro-wind-solar hybrid system in step (2.1), the remaining output after deducting the wind-solar output is allocated among the cascades. The K value discrimination method is used to determine the water supply and storage order of the cascades. When storing water, the power station with a large K value stores water first; when supplying water, the power station with a small K value supplies water first. The discriminant formula K of power station n is n The value calculation is as follows:

[0026]

[0027] Where: ΔE x Indicates the water storage ΔV due to power station n n This results in an increase in energy in the power station n and the downstream hydropower station; is the increase in water energy due to the increase in water head at power station n, W n is the current storage capacity of power station n; The downstream hydropower station is the water storage ΔV due to this power station n n The increased energy, represents the set of downstream power stations of power station n; ΔV n is the unit water storage capacity of power station n; η k is the average water consumption rate of power station k from the initial water level to the water storage level; ΔE z Indicates the water storage ΔV of power station n n resulting in increased energy from upstream hydroelectric power stations; V represents the set of upstream power stations of power station n; k represents the water storage above the dead storage capacity of the upstream power station k, W k is the interval water inflow of power station k; The power station n has a storage capacity of V n When , the relationship function between the change of unit storage capacity and the change of water consumption rate;

[0028] (3) Use fuzzy theory to characterize the high-dimensional and multiple uncertain probabilities of water, wind and light;

[0029] (3.1) Assuming that the runoff and wind and solar power output prediction errors are fuzzy variables and the prediction error distribution conforms to the Cauchy distribution, the membership function expression of the runoff and wind and solar power output prediction error ε is:

[0030]

[0031] Where: EP and EN are the statistical averages of positive and negative errors in the uncertainty set of runoff and wind and solar output, and σ is the weight;

[0032] (3.2) Use Python programming language to import long-term historical forecast and measured data of runoff and wind and solar power output from Excel files, and use Python-Scipy library-cauchy.fit function to fit the parameters of Cauchy distribution to the calculated forecast error data;

[0033] (3.3) To express the comprehensive membership between the runoff and wind and solar power output of each power station in the same period and between the runoff and wind and solar power output of each power station in different periods, the following two types of fuzzy relationships are defined:

[0034] The fuzzy relationship between runoff and wind and solar output of each power station in the same period is as follows:

[0035] h1(Q1 t ,Q2 t ,...,Qn t ,Pwp t )=min{f(Q1 t ),f(Q2 t ),...,f(Qn t ),f(Pwp t )} (11)

[0036] Where: Q1 t ,Q2 t ,...,Qn t are the inflow or interval flow of power stations 1, 2, ..., n in period t; Pwp t is the output of wind and solar power in period t; f(·) is the corresponding membership function;

[0037] The fuzzy relationship between runoff and wind and solar output of each power station at different time periods is expressed as follows:

[0038]

[0039] h3(Pwp1,Pwp2,...,Pwp T )=min{f(Pwp1),f(Pwp2),...,f(Pwp T )} (13)

[0040] Where: h2(·) and h3(·) are the fuzzy relationships between runoff in different time periods and wind and solar output in different time periods, i.e., the comprehensive membership; is the runoff of power station n in the period 1,2,…,T; Pwp1,Pwp2,…,Pwp T is the output of wind and solar power in the period 1, 2, …, T;

[0041] (4) Select key risk indicators for flood and dry seasons and establish a key risk indicator set for flood and dry seasons of the water-wind-solar complementary system, as follows:

[0042] (4.1) Risk of power shortage during dry season R s for

[0043]

[0044] Where: P i shortage is the risk characterization value of insufficient output in scenario i, where the risk is 1 and the risk is 0; P i,t is the system output value of the tth period of scenario i; I is the total number of simulated scenarios, T1 is the total number of dry period periods, P f It is the guaranteed output of the water-wind-solar complementary system;

[0045] (4.2) Risk of water abandonment during flood season R w for

[0046]

[0047] Where: S i is the amount of water abandoned in scenario i; Sd is the water abandonment control threshold; spill i,t is the amount of water abandoned in the tth period of scenario i;

[0048] (4.3) Risk of insufficient energy storage at the end of the year R e for

[0049]

[0050] In the formula: Eend i is the year-end energy storage under scenario i; E min It is the system's minimum energy storage requirement at the end of the year, and energy storage below this is considered insufficient;

[0051] (4.4) Wind and solar power curtailment risk R c for

[0052]

[0053] curtail i,t =f(Ph i,t ) t∈T (21)

[0054] Where: C i is the wind and solar curtailment in scenario i; Cd is the curtailment control threshold of cascade power stations; curtail i,t is the wind and solar curtailment in the tth period of scenario i; f(·) is the curtailment function of the wind-hydro-solar hybrid system; Ph i,t is the hydropower output in the tth period of scenario i; T is the number of periods in the whole year;

[0055] The calculation method for the risk loss value of power shortage during dry season, water abandonment during flood season, insufficient energy storage at the end of the year and wind and solar power abandonment under each indicator scenario is as follows:

[0056]

[0057] Where: is the index loss value of power shortage in dry season, water abandonment in flood season, insufficient energy storage at the end of the year and wind and solar power abandonment in the water-wind-solar complementary system; μ i,t is the fuzzy membership of scene i at time period t; μ i is the comprehensive fuzzy membership of the i-th scene;

[0058] (5) Quantification of risks of power shortage during dry season, water abandonment during flood season, wind and solar power abandonment, and insufficient energy storage at the end of the year

[0059] The runoff prediction error and wind and solar power output prediction error are recorded as fuzzy variables Without considering the difference between wind and solar power, wind power and photovoltaic power are collectively referred to as new energy. The steps to quantify the risks are as follows:

[0060] (5.1) Based on the historical runoff and renewable energy output forecast and measured data, the parameters of the runoff and wind and solar output membership functions are calculated using the Python-Scipy library-cauchy.fit function, and then the fuzzy membership functions of the runoff prediction error and renewable energy output prediction error are determined respectively. z represents the uncertain variable of runoff or wind and solar output, and m represents the month;

[0061] (5.2) Based on the error membership function of inflow and renewable energy output, a series of real numbers are randomly generated using the Monte Carlo simulation method in the Python-random package. and the corresponding membership N is the number of random numbers;

[0062] (5.3) Perform simulation operation; in the dry season, starting from the beginning of the year, according to the dry season drawdown plan, a fixed water level calculation is used to perform monthly simulation; for any month m, a series of According to the initial and final water levels in the scheme, a series of complementary system total outputs Ps are obtained: m,k Wind and solar power curtailment during dry season And determine the degree of membership If the final water level in the scheme is not feasible, the simulated water level is used instead of the final water level in the scheme for subsequent simulation calculations; during the flood season, a five-stage dispatching rule is used for monthly simulation; starting from the beginning of the flood season, a series of For any scenario Get the year-end energy storage Eend k , wind and solar power abandonment during flood season and flood season wastewater k And the corresponding membership ml represents the starting month of the flood season. and Add them together to get the annual wind and solar power curtailment under scenario k;

[0063] (5.4) According to step (3), calculate the risk probability and risk loss of power shortage during dry season, water abandonment during flood season, insufficient energy storage at the end of the year and wind and solar power abandonment indicators;

[0064] (5.5) For other cascade pre-flood energy storage, repeat (5.2)-(5.6) to finally obtain the refined relationship between cascade energy storage and power shortage during dry season, water abandonment during flood season, insufficient energy storage at the end of the year and the risk of wind and solar power abandonment.

[0065] Compared with the existing methods, the beneficial effects of the present invention are as follows: it is possible to obtain a refined relationship between the cascade pre-flood energy storage of the water, wind and solar systems and the multi-dimensional operating risks and system benefits, which is beneficial for dispatching personnel to accurately grasp the possible risks and potential maximum losses faced by the cascade pre-flood energy storage decision-making under multiple uncertainties of future water, wind and solar systems, thereby conducting multi-dimensional risk control and obtaining a cascade pre-flood energy storage control plan that balances risks and takes into account benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 This is a schematic diagram of the five-stage dispatching rules of the water-wind-solar complementary system;

[0067] Figure 2 It is a VaR diagram of risk loss distribution;

[0068] Figure 3 is the relationship between pre-flood energy storage and various risk probabilities; (1) is the risk of power shortage during the dry season; (2) is the risk of insufficient energy storage at the end of the year; (3) is the risk of water abandonment during the flood season; (4) is the risk of wind and solar power abandonment;

[0069] Figure 4 is the distribution of power shortage losses during the dry season and the risk value; where (1) is a normal water year with 7 energy storage before the flood season; (2) is a normal water year with 8 energy storage before the flood season; (3) is a dry year with 9 energy storage before the flood season; (4) is a dry year with 4 energy storage before the flood season; (5) is a dry year with 5 energy storage before the flood season; (6) is a dry year with 6 energy storage before the flood season; (7) is a dry year with 7 energy storage before the flood season; (8) is a dry year with 8 energy storage before the flood season; (9) is a dry year with 9 energy storage before the flood season;

[0070] Figure 5 is the system power generation under different pre-flood energy storage conditions; where (1) is a flood year; (2) is a normal water year; and (3) is a dry year. DETAILED DESCRIPTION

[0071] The specific implementation of the present invention is further described below in conjunction with the accompanying drawings and technical solutions.

[0072] A method for analyzing and controlling the risk of cascade energy storage before flood season in a water-wind-solar complementary system comprises the following steps:

[0073] (1) Taking the cascade pre-flood energy storage, monthly power control during the dry season, and conventional restrictions on hydropower dispatch as constraints, a dry season drawdown optimization model is constructed to obtain the optimal dry season drawdown scheme. The objective function of the dry season drawdown optimization model is as follows:

[0074]

[0075] Where E is the expected power generation under each scenario; J, T1, N represent the runoff uncertainty scenario, the number of dry season scheduling periods and the set of power stations, respectively, and j, t, n are the corresponding set elements; P j is the jth scene probability; Ph j,n,t is the output of power station n in the tth period of scenario j; Pwp j,t is the wind and solar power output in the tth period of scene j; Δt is the corresponding duration of the period (s);

[0076] The energy storage constraints before the flood season and the monthly power control constraints during the dry season are as follows:

[0077] 1) Constraints on cascade energy storage before flood season

[0078]

[0079] Where: To store energy before the flood season for power station n; E tar To store energy for the target before the flood season; Refers to the storage capacity of power station n at the end of the period; Refers to the total storage capacity of the power stations upstream of power station n, is the set of upstream power stations of power station n; η n is the average water consumption rate of power station n;

[0080] 2) Electricity control constraints during dry season months

[0081]

[0082] Where: K t is the control ratio of power generation in period t to dry period (historical statistics); ξ is the control error;

[0083] Using Gurobi solver as the modeling and solving platform, Python programming and pyomo modeling language are used to linearize the nonlinear constraints in the dry season drawdown optimization model and convert it into mixed integer linear programming for solving;

[0084] (2) Construct water storage scheduling rules, including five-stage hedging scheduling rules and cascade output allocation method, and the cascade output allocation method adopts K value discrimination method;

[0085] (2.1) Establish hedging dispatching rules: Use Python-pyomo modeling program to establish parameter linear optimization method to construct five-stage hedging dispatching rules to determine the monthly output of the water-wind-solar complementary system. The principle is shown in Figure 1 . In the figure, OAGD is the three-stage dispatching rule of the water-wind-solar complementary system, in which OA is the output section below the guaranteed output; AG is the output guarantee section; and GD is the increased output section. The hedging dispatching rule of the present invention is based on the three-stage dispatching rule, and adds points B and C between the AG and GD sections respectively, to establish the hedging dispatching rule OABCD containing the BC hedging section, in which the horizontal coordinates of the intersection of the BC section with AG and GD are parameters a and b. The method for determining parameters a and b is optimized based on the historical output results of water, wind and solar power, and the specific expressions are as follows:

[0086]

[0087] Where: k and d are the slope and intercept of the BC segment curve; E t is the average available energy of historical hydropower and wind power in period t; T2 is the number of scheduling periods during the water storage adjustment period; P t is the historical average output of the corresponding period t; since BC intersects with the AB segment and the GD segment, the expressions of the AB segment and the GD segment are combined, and a and b are calculated by the following formula:

[0088]

[0089] Where: P M It is the maximum output of the water-wind-solar complementary system; P f It is the guaranteed output of the water-wind-solar complementary system; is the fully available storage capacity of the cascade power station (excluding dead storage capacity); Δm is the number of hours in a month;

[0090] (2.2) Cascade output allocation: After obtaining the monthly output of the hydro-wind-solar hybrid system in step (2.1), the remaining output after deducting the wind-solar output is allocated among the cascades. The K value discrimination method is used to determine the water supply and storage order of the cascades. When storing water, the power station with a large K value stores water first; when supplying water, the power station with a small K value supplies water first. The discriminant formula K of power station n is n The value calculation is as follows:

[0091]

[0092] Where: ΔE x Indicates the water storage ΔV due to power station n n This results in an increase in energy in the power station n and the downstream hydropower station; is the increase in water energy due to the increase in water head at power station n, W n is the current storage capacity of power station n; The downstream hydropower station is the water storage ΔV due to this power station n n The increased energy, represents the set of downstream power stations of power station n; ΔV n is the unit water storage capacity of power station n; η k is the average water consumption rate of power station k from the initial water level to the water storage level; ΔE z Indicates the water storage ΔV of power station n n resulting in increased energy from upstream hydroelectric power stations; V represents the set of upstream power stations of power station n; k represents the water storage above the dead storage capacity of the upstream power station k, W k is the interval water inflow of power station k; The power station n has a storage capacity of V n When , the relationship function between the change of unit storage capacity and the change of water consumption rate;

[0093] (3) Use fuzzy theory to characterize the high-dimensional and multiple uncertain probabilities of water, wind and light;

[0094] (3.1) Assuming that the runoff and wind and solar power output prediction errors are fuzzy variables and the prediction error distribution conforms to the Cauchy distribution, the membership function expression of the runoff and wind and solar power output prediction error ε is:

[0095]

[0096] Where: EP and EN are the statistical averages of positive and negative errors in the uncertainty set of runoff and wind and solar output, and σ is the weight;

[0097] (3.2) Use Python programming language to import long-term historical forecast and measured data of runoff and wind and solar power output from Excel files, and use Python-Scipy library-cauchy.fit function to fit the parameters of Cauchy distribution to the calculated forecast error data;

[0098] (3.3) To express the comprehensive membership between the runoff and wind and solar power output of each power station in the same period and between the runoff and wind and solar power output of each power station in different periods, the following two types of fuzzy relationships are defined:

[0099] The fuzzy relationship between runoff and wind and solar output of each power station in the same period is as follows:

[0100] h1(Q1 t ,Q2 t ,...,Qn t ,Pwp t )=min{f(Q1 t ),f(Q2 t ),...,f(Qn t ),f(Pwp t )} (11)

[0101] Where: Q1 t ,Q2 t ,...,Qn t are the inflow or interval flow of power stations 1, 2, ..., n in period t; Pwp t is the output of wind and solar power in period t; f(·) is the corresponding membership function;

[0102] The fuzzy relationship between runoff and wind and solar output of each power station at different time periods is expressed as follows:

[0103]

[0104] h3(Pwp1,Pwp2,...,Pwp T )=min{f(Pwp1),f(Pwp2),...,f(Pwp T )} (13)

[0105] Where: h2(·) and h3(·) are the fuzzy relationships between runoff in different time periods and wind and solar output in different time periods, i.e., the comprehensive membership; is the runoff of power station n in the period 1,2,…,T; Pwp1,Pwp2,…,Pwp T is the output of wind and solar power in the period 1, 2, …, T;

[0106] (4) Select key risk indicators for flood and dry seasons and establish a key risk indicator set for flood and dry seasons of the water-wind-solar complementary system, as follows:

[0107] (4.1) Risk of power shortage during dry season R s for

[0108]

[0109] Where: P i shortage is the risk characterization value of insufficient output in scenario i, where the risk is 1 and the risk is 0; P i,t is the system output value of the tth period of scenario i; I is the total number of simulated scenarios, T1 is the total number of dry period periods, P fIt is the guaranteed output of the water-wind-solar complementary system;

[0110] (4.2) Risk of water abandonment during flood season R w for

[0111]

[0112] Where: S i is the amount of water abandoned in scenario i; Sd is the water abandonment control threshold; spill i,t is the amount of water abandoned in the tth period of scenario i;

[0113] (4.3) Risk of insufficient energy storage at the end of the year R e for

[0114]

[0115] In the formula: Eend i is the year-end energy storage under scenario i; E min It is the system's minimum energy storage requirement at the end of the year, and energy storage below this is considered insufficient;

[0116] (4.4) Wind and solar power curtailment risk R c for

[0117]

[0118] curtail i,t =f(Ph i,t ) t∈T (21)

[0119] Where: C i is the wind and solar curtailment in scenario i; Cd is the curtailment control threshold of cascade power stations; curtail i,t is the wind and solar curtailment in the tth period of scenario i; f(·) is the curtailment function of the wind-hydro-solar hybrid system; Ph i,t is the hydropower output in the tth period of scenario i; T is the number of periods in the whole year;

[0120] The calculation method for the risk loss value of power shortage during dry season, water abandonment during flood season, insufficient energy storage at the end of the year and wind and solar power abandonment under each indicator scenario is as follows:

[0121]

[0122] Where: is the index loss value of power shortage in dry season, water abandonment in flood season, insufficient energy storage at the end of the year and wind and solar power abandonment in the water-wind-solar complementary system; μ i,t is the fuzzy membership of scene i at time period t; μ i is the comprehensive fuzzy membership of the i-th scene;

[0123] (5) Quantification of risks of power shortage during dry season, water abandonment during flood season, wind and solar power abandonment, and insufficient energy storage at the end of the year

[0124] The runoff prediction error and wind and solar power output prediction error are recorded as fuzzy variables Without considering the difference between wind and solar power, wind power and photovoltaic power are collectively referred to as new energy. The steps to quantify the risks are as follows:

[0125] (5.1) Based on the historical runoff and renewable energy output forecast and measured data, the parameters of the runoff and wind and solar output membership functions are calculated using the Python-Scipy library-cauchy.fit function, and then the fuzzy membership functions of the runoff prediction error and renewable energy output prediction error are determined respectively. z represents the uncertain variable of runoff or wind and solar output, and m represents the month;

[0126] (5.2) Based on the error membership function of inflow and renewable energy output, a series of real numbers are randomly generated using the Monte Carlo simulation method in the Python-random package. and the corresponding membership N is the number of random numbers;

[0127] (5.3) Perform simulation operation; in the dry season, starting from the beginning of the year, according to the dry season drawdown plan, a fixed water level calculation is used to perform monthly simulation; for any month m, a series of According to the initial and final water levels in the scheme, a series of complementary system total outputs Ps are obtained: m,k Wind and solar power curtailment during dry season And determine the degree of membership If the final water level in the scheme is not feasible, the simulated water level is used instead of the final water level in the scheme for subsequent simulation calculations; during the flood season, a five-stage dispatching rule is used for monthly simulation; starting from the beginning of the flood season, a series of For any scenario Get the year-end energy storage Eend k , wind and solar power abandonment during flood season and flood season wastewater k And the corresponding membership ml represents the starting month of the flood season. and Add them together to get the annual wind and solar power curtailment under scenario k;

[0128] (5.4) According to step (3), calculate the risk probability and risk loss of power shortage during dry season, water abandonment during flood season, insufficient energy storage at the end of the year and wind and solar power abandonment indicators;

[0129] (5.5) For other cascade pre-flood energy storage, repeat (5.2)-(5.6) to finally obtain the refined relationship between cascade energy storage and power shortage during dry season, water abandonment during flood season, insufficient energy storage at the end of the year and the risk of wind and solar power abandonment.

[0130] The present invention is verified by taking the water-wind-solar complementary base in the downstream of a particularly large river basin as an example. The cascade power stations use four reservoirs, XW, MW, NZD and JH. According to the plan, wind power and photovoltaic power are set to 570MW and 4497MW respectively. Three typical level years of flood, normal and drought were selected as future years with different frequencies to test the effectiveness of the present invention. Since the regulation capabilities of MW and JH are relatively weak, the water levels of XW and NZD at the beginning of the year are set to 1230m and 805m according to the historical operation data under the long-term given monthly water level operation. According to the flood and drought variation law in the downstream of the river basin, July is selected as the pre-flood energy storage month. Referring to the historical dispatching and operating water level process of XW and NZD, the minimum drawdown water level interval of XW is set to [1166,1186.5], and that of NZD is set to [765,797.6]. According to calculations, the minimum pre-flood energy storage of the cascade is 13.15 billion kWh and the maximum is 20.13 billion kWh. Within the energy storage range, 9 typical pre-flood energy storages are evenly discretized with a step size of 880 million kWh for risk-benefit analysis, and the energy storages are numbered 1 to 9. The minimum energy storage demand of the cascade at the end of the year is 22.13 billion kWh. According to relevant policies and historical operating experience, the thresholds for water and electricity abandonment are set at 6% and 10%.

[0131] Figure 3 The relationship between the cascade pre-flood energy storage and the probability of each risk is shown. It can be seen that there is almost no year-end energy storage shortage and power shortage during the dry season in the wet year, but there is a greater risk of water abandonment and wind and solar power abandonment. Under the highest pre-flood energy storage, the risks of water abandonment and wind and solar power abandonment are 0.54 and 0.84 respectively; on the contrary, there is no risk of water abandonment and wind and solar power abandonment in the dry year, but the energy storage is too high, the water volume is insufficient during the dry season, and there is a greater risk of power shortage. The risk probability corresponding to the maximum pre-flood energy storage is 1. In addition, if the pre-flood energy storage is too low, the water storage during the flood season in the dry year is insufficient, and the power station often cannot meet the minimum year-end energy storage. The energy storage risks corresponding to pre-flood energy storage 1, 2 and 3 are 0.306, 0.175 and 0.095. In the normal water year, the year-end energy storage can almost meet the requirements and the risk can be ignored. However, it should be avoided that the pre-flood energy storage is too high. The energy storage of 7, 8 and 9 will bring a considerable risk of power shortage during the dry season, which are 0.085, 0.204 and 0.52 respectively. As the pre-flood energy storage increases, the water abandonment during the flood season increases significantly, especially when the energy storage exceeds 5. In comparison, the risk of power abandonment increases with the increase of pre-flood energy storage but is not significant. The risk ranges of the two are [0.02, 0.34] and [0.04, 0.15], respectively.

[0132] according to Figure 3The risk relationship diagram shows that decision makers can make decisions under certain risk tolerance. For example, assuming the tolerance is 20%, in a flood year, the pre-flood energy storage should be controlled below 4 as much as possible to avoid excessive water and power abandonment risks; in a normal water year, it can be appropriately raised, but it should not exceed 6 to avoid excessive water abandonment risks; in a dry year, a trade-off can be made between 2-5. In addition, the comprehensive determination of pre-flood energy storage can also refer to the subsequent risk loss and benefit analysis.

[0133] We further calculated the risk loss distribution and risk value (confidence level 0.95) for risk probabilities greater than 0.05. Figure 4 As can be seen from Table 1, under the same pre-flood energy storage, the greater the water inflow, the lower the risk value of power shortage in the dry season, and the higher the risk value of water abandonment in the flood season. In addition, the greater the risk, the corresponding risk loss tends to increase accordingly, that is, the greater the risk value. Figure 4 Taking the power outage loss during the dry season in a dry year as an example, the maximum output loss corresponding to energy storage 4 is 42.5MW, and the maximum output loss corresponding to energy storage 9 is 423.6MW, which is nearly ten times the difference. However, the relationship between risk probability and risk value is not linearly positively correlated. For example, in a flood year, the probability of power abandonment of energy storage 5 to 9 before the flood season differs by 0.12 (see Figure 1 ), but the risk values ​​are almost equal (see Table 2). This inconsistent relationship can give decision makers a certain decision space, and further illustrates the significance and necessity of comprehensively characterizing risks with risk probability and risk loss, confirming the effectiveness and comprehensiveness of the risk quantification of the present invention.

[0134] Table 1 Risk value of abandoned water

[0135]

[0136] Table 2 Risk value of power curtailment

[0137]

[0138]

[0139] Figure 5 is the system power generation under different pre-flood energy storage. With the increase of pre-flood energy storage, the water consumption in the dry season decreases, the water storage in the flood season decreases, and the power generation head and water consumption for power generation increase. Therefore, the power generation in the dry season decreases with the increase of pre-flood energy storage, and the power generation in the flood season increases with the increase of pre-flood energy storage. The overall power generation in wet years and dry years decreases with the increase of pre-flood energy storage. This is because, in wet years, excessive pre-flood energy storage increases the loss of abandoned water and reduces the power generation efficiency; while in dry years, the lower the pre-flood energy storage, the greater the system water consumption for power generation and the greater the power generation efficiency. Benefit analysis can make up for the problem of insufficient information of pre-flood energy storage that only considers risk analysis. For example, as mentioned above, the pre-flood energy storage in normal water years needs to be controlled below the energy storage of 6 as much as possible, from Figure 5It can be seen that the energy storage capacity of 5 is the highest and the risk value is relatively small (see Table 1 and Table 2), and the difference is not big. Therefore, the energy storage capacity before the flood season should be controlled at 5.

[0140] In order to further test the effectiveness of the present invention, the long series of system benefit and risk results were simulated and compared with other methods. The comparison method adopted the commonly used long-term random expectation scheduling model. First, with the benefit and risk analysis results as a reference, the decision of the dispatcher on pre-flood energy storage was simulated, and then the cascade pre-flood energy storage was obtained by solving the comparison model. The results of the two are shown in Table 3.

[0141] Table 3 Pre-flood energy storage decision-making plan

[0142]

[0143] Based on the determined pre-flood energy storage, the long-term scheduling process of the two was simulated. The uniform drawdown rule is adopted for operation during the drawdown period, and the water storage scheduling rule proposed above is adopted for operation during the water storage adjustment period. After many years of simulation, the index results of the two are shown in Table 4. It can be seen that the pre-flood energy storage control screened by the present invention has better performance in power generation, power shortage during dry season, wind and solar power abandonment, and water abandonment during flood season, with an average reduction of 42% in risk loss. This is because the pre-flood energy storage selected by the present invention avoids the above-mentioned risks and performs better at the multi-year level, which verifies the effectiveness and practicality of the present invention in actual scheduling work.

[0144] Table 4 Comparison of simulation indicators

[0145]

Claims

1. A risk analysis and control method for pre-flood energy storage in a cascade water-wind-solar hybrid system, characterized in that: The steps include: (1) Taking the cascade pre-flood energy storage, monthly power control during the dry season, and conventional restrictions on hydropower dispatch as constraints, a dry season drawdown optimization model is constructed to obtain the optimal dry season drawdown scheme. The objective function of the dry season drawdown optimization model is as follows: Where E is the expected power generation under each scenario; J, T1, N represent the runoff uncertainty scenario, the total number of dry seasons and the set of power stations, respectively, and j, t, n are the corresponding set elements; P j is the jth scene probability; Ph j,n,t is the output of power station n in the tth period of scenario j; Pwp j,t is the wind and solar power output in the tth period of scene j; Δt is the corresponding duration of the period (s); The energy storage constraints before the flood season and the monthly power control constraints during the dry season are as follows: 1) Constraints on cascade energy storage before flood season Where: To store energy before the flood season for power station n; E tar To store energy for the target before the flood season; Refers to the storage capacity of power station n at the end of the period; Refers to the total storage capacity of the power stations upstream of power station n, is the set of upstream power stations of power station n; η n is the average water consumption rate of power station n; 2) Electricity control constraints during dry season months Where: K t is the control ratio of power generation in period t to dry period; ξ is the control error; Using Gurobi solver as the modeling and solving platform, Python programming and pyomo modeling language are used to linearize the nonlinear constraints in the dry season drawdown optimization model and convert it into mixed integer linear programming for solving; (2) Construct water storage scheduling rules, including five-stage hedging scheduling rules and cascade output allocation method, and the cascade output allocation method adopts K value discrimination method; (2.1) Establish hedging dispatching rules: The Python-pyomo modeling program is used to establish a parameter linear optimization method to construct a five-stage hedging dispatching rule to determine the monthly output of the water-wind-solar complementary system; among them, OAGD is a three-stage dispatching rule for the water-wind-solar complementary system, in which the OA segment is the output segment below the guaranteed output, the AG segment is the guaranteed output segment, and the GD segment is the increased output segment; on the basis of the three-stage dispatching rule, point B and point C are added between the AG segment and the GD segment respectively, and the hedging dispatching rule OABCD containing the BC hedging segment is established, in which the horizontal coordinates of the intersection of the BC segment with the AG segment and the GD segment are parameters a and b; the method for determining parameters a and b is optimized according to the historical output results of the water-wind-solar system, and the specific expressions are as follows: Where: k and d are the slope and intercept of the BC segment curve; E t is the average available energy of historical hydropower and wind power in period t; T2 is the number of scheduling periods during the water storage adjustment period; P t is the historical average output of the corresponding period t; since BC intersects with the AB segment and the GD segment, the expressions of the AB segment and the GD segment are combined, and a and b are calculated by the following formula: Where: P f It is the guaranteed output of the water-wind-solar complementary system; is the fully available energy storage of the cascade power station; Δm is the number of hours in a month; (2.2) Cascade output allocation: After obtaining the monthly output of the hydro-wind-solar hybrid system in step (2.1), the remaining output after deducting the wind-solar output is allocated among the cascades. The K value discrimination method is used to determine the water supply and storage order of the cascades. When storing water, the power station with a large K value stores water first; when supplying water, the power station with a small K value supplies water first. The discriminant formula K of power station n is n The value calculation is as follows: Where: ΔE x Indicates the water storage ΔV due to power station n n This results in an increase in energy in the power station n and the downstream hydropower station; is the increase in water energy due to the increase in water head at power station n, W n is the current storage capacity of power station n; The downstream hydropower station is the water storage ΔV due to this power station n n The increased energy, represents the set of downstream power stations of power station n; ΔV n is the unit water storage capacity of power station n; η k is the average water consumption rate of power station k from the initial water level to the water storage level; ΔE z Indicates the water storage ΔV of power station n n resulting in increased energy from upstream hydroelectric power stations; V represents the set of upstream power stations of power station n; k represents the water storage above the dead storage capacity of the upstream power station k, W k is the interval water inflow of power station k; The power station n has a storage capacity of V n When , the relationship function between the change of unit storage capacity and the change of water consumption rate; (3) Use fuzzy theory to characterize the high-dimensional and multiple uncertain probabilities of water, wind and light; (3.1) Assuming that the runoff and wind and solar power output prediction errors are fuzzy variables and the prediction error distribution conforms to the Cauchy distribution, the membership function expression of the runoff and wind and solar power output prediction error ε is: Where: EP and EN are the statistical averages of positive and negative errors in the uncertainty set of runoff and wind and solar output, and σ is the weight; (3.2) Use Python programming language to import long-term historical forecast and measured data of runoff and wind and solar power output from Excel files, and use Python-Scipy library-cauchy.fit function to fit the parameters of Cauchy distribution to the calculated forecast error data; (3.3) To express the comprehensive membership between the runoff and wind and solar power output of each power station in the same period and between the runoff and wind and solar power output of each power station in different periods, the following two types of fuzzy relationships are defined: The fuzzy relationship between runoff and wind and solar output of each power station in the same period is as follows: h1(Q1 t ,Q2 t ,...,Qn t ,Pwp t )=min{f(Q1 t ),f(Q2 t ),...,f(Qn t ),f(Pwp t )} (11) Where: Q1 t ,Q2 t ,...,Qn t are the inflow or interval flow of power stations 1, 2, ..., n in period t; Pwp t is the output of wind and solar power in period t; f(·) is the corresponding membership function; The fuzzy relationship between runoff and wind and solar output of each power station at different time periods is expressed as follows: h3(Pwp1,Pwp2,...,Pwp T )=min{f(Pwp1),f(Pwp2),...,f(Pwp T )} (13) Where: h2(·) and h3(·) are the fuzzy relationships between runoff in different time periods and wind and solar output in different time periods, i.e., the comprehensive membership; is the runoff of power station n in the period 1,2,…,T; Pwp1,Pwp2,…,Pwp T is the output of wind and solar power in the period 1, 2, …, T; (4) Select key risk indicators for flood and dry seasons and establish a key risk indicator set for flood and dry seasons of the water-wind-solar complementary system, as follows: (4.1) Risk of power shortage during dry season R s for Where: P i shortage is the risk characterization value of insufficient output in scenario i, where the risk is 1 and the risk is 0; P i,t is the system output value of the tth period of scenario i; I is the total number of simulated scenarios, T1 is the total number of dry period periods, P f It is the guaranteed output of the water-wind-solar complementary system; (4.2) Risk of water abandonment during flood season R w for Where: S i is the amount of water abandoned in scenario i; Sd is the water abandonment control threshold; spill i,t is the amount of water abandoned in the tth period of scenario i; (4.3) Risk of insufficient energy storage at the end of the year R e for In the formula: Eend i is the year-end energy storage under scenario i; E min It is the system's minimum energy storage requirement at the end of the year, and energy storage below this is considered insufficient; (4.4) Wind and solar power curtailment risk R c for curtail i,t =f(Ph i,t ) t∈T (21) Where: C i is the wind and solar curtailment in scenario i; Cd is the curtailment control threshold of cascade power stations; curtail i,t is the wind and solar curtailment in the tth period of scenario i; f(·) is the curtailment function of the wind-hydro-solar hybrid system; Ph i,t is the hydropower output in the tth period of scenario i; T is the number of periods in the whole year; The calculation method for the risk loss value of power shortage during dry season, water abandonment during flood season, insufficient energy storage at the end of the year and wind and solar power abandonment under each indicator scenario is as follows: Where: μ is the index loss value of power shortage in dry season, water abandonment in flood season, insufficient energy storage at the end of the year and wind and solar power abandonment in the water-wind-solar complementary system; i,t is the fuzzy membership of scene i at time period t; μ i is the comprehensive fuzzy membership of the i-th scene; (5) Quantification of risks of power shortage during dry season, water abandonment during flood season, wind and solar power abandonment, and insufficient energy storage at the end of the year The runoff prediction error and wind and solar power output prediction error are recorded as fuzzy variables Without considering the difference between wind and solar power, wind power and photovoltaic power are collectively referred to as new energy. The steps to quantify the risks are as follows: (5.1) Based on the historical runoff and renewable energy output forecast and measured data, the parameters of the runoff and wind and solar output membership functions are calculated using the Python-Scipy library-cauchy.fit function, and then the fuzzy membership functions of the runoff prediction error and renewable energy output prediction error are determined respectively. 1≤m≤12, 1≤z≤Z, z represents the uncertain variable of runoff or wind and solar power output, and m represents the month; (5.2) Based on the error membership function of inflow and renewable energy output, a series of real numbers are randomly generated using the Monte Carlo simulation method in the Python-random package. and the corresponding membership N is the number of random numbers; (5.3) Perform simulation operation; in the dry season, starting from the beginning of the year, according to the dry season drawdown plan, a fixed water level calculation is used to perform monthly simulation; for any month m, a series of According to the initial and final water levels in the scheme, a series of complementary system total outputs Ps are obtained: m,k Wind and solar power curtailment during dry season And determine the degree of membership If the final water level in the scheme is not feasible, the simulated water level is used instead of the final water level in the scheme for subsequent simulation calculations; during the flood season, a five-stage dispatching rule is used for monthly simulation; starting from the beginning of the flood season, a series of For any scenario Get the year-end energy storage Eend k , wind and solar power abandonment during flood season and flood season wastewater k And the corresponding membership ml represents the starting month of the flood season. and Add them together to get the annual wind and solar power curtailment under scenario k; (5.4) According to step (4), calculate the risk probability and risk loss of power shortage during dry season, water abandonment during flood season, insufficient energy storage at the end of the year and wind and solar power abandonment indicators; (5.5) For other cascade pre-flood energy storage, repeat (5.2)-(5.4) to finally obtain the refined relationship between cascade energy storage and power shortage during dry season, water abandonment during flood season, insufficient energy storage at the end of the year and the risk of wind and solar power abandonment.

Images