A random optimization scheduling method and system for determining wind power operation backup

Abstract

A random optimization scheduling method and system for determining wind power operation reserve, comprising: collecting actual operation data and prediction data of wind turbines, and establishing an error probability distribution function of intra-day wind power prediction based on kernel density estimation; determining the credibility of intra-day wind power prediction by using the inverse function of the error probability distribution function and the set prediction power accuracy probability; taking the output of the wind turbine greater than the credibility as the operation reserve capacity provided by the wind power; using a full-scenario feasible random optimization method to screen out a key scenario set from the intra-day wind power output scenario set; for the key scenario set, establishing a unit commitment model for wind power into operation reserve with the minimum operation cost of the power system as the objective function; in the intra-day operation stage, determining the operation reserve capacity of the synchronous thermal power unit and the wind turbine by using the unit commitment model for wind power into operation reserve. The unexpectedness of economic dispatching decision and the full-scenario feasibility of unit commitment decision are ensured.

Description

Technical Field

[0001] This invention belongs to the field of new energy technology, specifically relating to a stochastic optimization scheduling method and system for determining wind power operation reserves. Background Technology

[0002] As a major form of new energy power generation, wind power's volatility and intermittency pose a severe challenge to the safe operation of the power system. To cope with the uncertainty of wind power, power systems with large-scale wind power grid connection need to reserve sufficient spinning reserve capacity for thermal power units to ensure the safe operation of the system.

[0003] In existing technologies, before a high proportion of wind power is integrated into the grid, reserve demand assessment employs deterministic methods, commonly using the maximum unit capacity within the system (N-1 criterion), a certain percentage of the system's peak load, or a combination of both. These methods convert uncertainties into certainties or percentages through established criteria, thereby assessing the system's reserve demand. However, after a high proportion of wind power is integrated into the grid, the relationship between system uncertainty and reserve demand becomes more complex, rendering deterministic methods inapplicable. It is now necessary to consider the uncertainty of wind power forecasting.

[0004] For short-term operational reserves on the generation side, in traditional power systems, synchronous generators mainly provide the positive and negative reserves required by the system. When the proportion of wind power connected to the grid is small, the dispatching department can reserve full reserves for wind power. In high-proportion renewable energy power systems, when wind power accounts for 10% of the total power consumption, the expected increase in reserve capacity is approximately 2% to 8% of the installed wind power capacity. The difference in required reserve capacity is due to the different number of hours in advance the wind power forecasting error is compensated for by the reserve capacity, resulting in different wind power forecasting error ranges and thus different increases in reserve demand. In this case, if thermal power units provide full reserve, it will cause a significant drop in the load factor of traditional thermal power units, leading to a significant increase in unit coal consumption rate and a significant increase in system reserve costs. Therefore, wind power can be included in the system's operational reserve as a usable generation reserve resource. Domestic and international scholars have conducted extensive research on the operational optimization of wind power in system reserves, with modeling wind power uncertainties being a core issue. Modeling methods utilizing wind power output range prediction information do not require precise probability density functions. Instead, they determine a series of scenarios that significantly impact system operation based on the upper and lower bounds of the prediction range. Representative methods include robust optimization and range optimization. Modeling methods utilizing wind power probabilistic prediction information employ probability density functions or cumulative distribution functions with precise mathematical expressions for prediction. Representative methods include scenario optimization, chance-constrained programming, and risk optimization. However, current research has not considered unexpected constraints and full-scenario feasibility. The second-stage decision in robust optimization is given under the condition that all random factors are known, relying on the realization values ​​of future uncertainties. Scenario-based stochastic optimization methods can only consider a finite number of key scenarios from a probability distribution perspective. Chance-constrained methods also use probabilistic constraints instead of traditional deterministic constraints to ensure the feasibility of scheduling schemes at a certain confidence level, but they cannot guarantee solution feasibility in the face of infinitely many scenarios in stochastic production simulations. Therefore, this paper proposes a stochastic unit combination model that incorporates wind power output into the reserve system, which can guarantee the unexpectedness of economic dispatch decisions and the full-scenario feasibility of unit combination decisions. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a stochastic optimization scheduling method and system for determining wind power operating reserves. It proposes incorporating wind turbine units as generation-side resources into the system's operating reserves and uses a feasible stochastic optimization method across all scenarios to determine the operating reserve capacity of wind power and traditional synchronous thermal power units in the power system, thereby improving the wind power absorption rate.

[0006] The present invention adopts the following technical solution.

[0007] A stochastic optimization scheduling method for determining wind power operating reserves includes:

[0008] Step 1: Collect actual operating data and forecast data of wind turbine units, and establish the error probability distribution function of intraday wind power forecast based on kernel density estimation;

[0009] Step 2: Using the inverse function of the error probability distribution function and the set prediction power accuracy probability, determine the credibility of the intraday wind power prediction power; use the output of wind turbine units that are greater than the credibility as the operating reserve capacity provided by wind power.

[0010] Step 3: Using the full-scenario feasible stochastic optimization method, select key scenario sets from the intraday wind power output scenario set; for the key scenario sets, establish a unit combination model for wind power to be included in the operating reserve with the objective function of minimizing the operating cost of the power system; during the intraday operation phase, use the unit combination model for wind power to be included in the operating reserve to determine the operating reserve capacity of synchronous thermal power units and wind power units.

[0011] Preferably, a Gaussian function is used as the kernel function, and the optimal bandwidth for kernel density estimation is:

[0012]

[0013] In the formula, h is the bandwidth of the kernel density estimation, n is the sample size, and σ is the standard deviation.

[0014] Preferably, the reliability of the intraday wind power forecast is as follows:

[0015] A p =G(1-C%)

[0016] In the formula,

[0017] A p To assess the reliability of intraday wind power forecasts,

[0018] G(·) is the inverse function of the probability distribution function of wind power prediction error.

[0019] C% represents the set probability of predicted power accuracy.

[0020] Preferably, the predicted power accuracy probability is set to 95%.

[0021] Preferably, the constraints on the operating reserve capacity provided by wind power are as follows:

[0022]

[0023] In the formula,

[0024] The operating reserve capacity provided for wind turbine j at time t.

[0025] w j,t Let be the predicted power of wind turbine j at time t.

[0026] Preferably, step 3 includes:

[0027] Step 3.1: Using wind turbine output as a random variable, a convex polyhedral uncertainty set U is used to describe the intraday wind power output scenario set; using the full-scenario feasible stochastic optimization method, the key scenario set selected from the intraday wind power output scenario set includes: basic scenario S BS Vertex Scene S SVS and extreme climbing scenario S ERS The key scenario set is represented as S = {S BS ,S SVS ,S ERS};

[0028] Step 3.2: For the key scenario set, with the objective function of minimizing the operating cost of the power system, establish a unit combination model for wind power to be included in the operating reserve; the unit combination model for wind power to be included in the operating reserve includes: a dispatch model and the constraints of the dispatch model;

[0029] Step 3.3: During the intraday operation phase, the operating reserve capacity of synchronous thermal power units and wind power units is determined using the unit combination model that includes wind power in the operating reserve.

[0030] Preferably, step 3.1 includes:

[0031] Step 3.1.1: Determine the basic scenario S based on historical operational data. BS In the process of randomly optimizing and generating scenarios, only one basic scenario is considered, namely the desired scenario, as follows:

[0032]

[0033] In the formula,

[0034] w BS This is a matrix representation of the wind turbine output in the basic scenario.

[0035] Let j be the output of wind turbine j at time t in the basic scenario.

[0036] K w The number of wind turbine units.

[0037] T represents the number of time periods.

[0038] E(·) is the expected value function.

[0039] w j,t The predicted output of wind turbine j at time t;

[0040] Step 3.1.2: Select vertices at the same sorting order m at each time step from the uncertain set U of the convex polyhedron to form the vertex scene S. SVS ,as follows:

[0041]

[0042] In the formula,

[0043] w SVS This represents the matrix representation of the wind turbine output in the vertex scenario.

[0044] Let m be the output of wind turbine j at time t in the vertex scenario, and let m be the sorting order.

[0045] N t Let U be the uncertainty set of a convex polyhedron at time t. t The number of vertices in the array;

[0046] Step 3.1.3 defines an odd-numbered scenario as one where wind power output has an upper limit during odd-numbered periods and a lower limit during even-numbered periods, and an even-numbered scenario as one where wind power output has a lower limit during odd-numbered periods and an upper limit during even-numbered periods. The odd-numbered and even-numbered scenarios together constitute the extreme ramping scenario S. ERS ;

[0047] S ERS The generation process is represented as follows:

[0048]

[0049]

[0050]

[0051] In the formula,

[0052] w OS This represents the matrix representation of wind turbine output in odd-number scenarios.

[0053] Let represent the output of wind turbine j at time t in the odd-number scenario.

[0054] w ES This represents the matrix representation of wind turbine output in even-number scenarios.

[0055] Let be the output of wind turbine j at time t in the even-number scenario.

[0056] T odd For odd-numbered moments,

[0057] T even For even-numbered moments,

[0058] These represent the upper and lower limits of the output of wind turbine j at time t, respectively.

[0059] Preferably, the method is generated using a full-scenario feasible stochastic optimization approach. A set of key scenarios, including: One vertex scene, one basic scene, and two extreme climbing scenes;

[0060] in,

[0061] Preferably, the scheduling model is as follows:

[0062]

[0063] In the formula,

[0064] C state For unit start-up and shutdown costs,

[0065] C cur For the cost of wind power abandonment,

[0066] C oper For coal consumption costs of thermal power units,

[0067] C back For system backup costs,

[0068] π s Let be the probability of the key scenario s.

[0069] Preferably, the costs satisfy the following relationship:

[0070]

[0071] In the formula,

[0072] a, b, and c are the coefficients of the quadratic, linear, and constant terms of coal consumption cost, respectively.

[0073] P g,t Let g be the output of the thermal power unit at time t.

[0074] These are the start-up cost and shutdown cost of thermal power unit g, respectively.

[0075] Let be the cost of wind power curtailment per unit of electricity at time t.

[0076] ΔR j,ｔ Let be the amount of wind power curtailed by wind turbine j at time t.

[0077] c ug c dgThese represent the purchase cost of the positive operating reserve capacity and the purchase cost of the negative operating reserve capacity of thermal power unit g, respectively.

[0078] R ｕｇ,ｔ R ｄｇ,t These represent the positive and negative operating reserve capacity of thermal power unit g at time t, respectively.

[0079] c wj The cost of purchasing the operating reserve capacity of wind turbine unit j.

[0080] The operating reserve capacity provided for wind turbine j at time t.

[0081] G represents the number of thermal power units.

[0082] K w The number of wind turbine units.

[0083] T represents the number of time periods.

[0084] Preferably, the constraints of the scheduling model include:

[0085] 1) Constraints between the operating state, start-up state, and shutdown state of thermal power units;

[0086] 2) Constraints on the minimum start-up and shutdown times of thermal power units:

[0087] 3) Constraints on the output of thermal power units;

[0088] 4) Constraints on wind turbine output;

[0089] 5) Constraints on the climbing ability of thermal power units;

[0090] 6) Constraints on system standby capacity.

[0091] Preferably, the constraints of the scheduling model also include: strongly unexpected constraints; as follows:

[0092]

[0093]

[0094]

[0095] In the formula,

[0096] These represent the odd-numbered scenarios S respectively. OS Even-numbered scenarios S ES The output of the thermal power unit g at time t.

[0097] x g,tLet g be the operating state of the thermal power unit at time t, and x be the operating state at startup. g,t =1, x when shutting down g,t =0,

[0098] S BS S SVS S ERS These are the basic scene, the vertex scene, and the extreme climbing scene.

[0099] P g,t Let g be the output of the thermal power unit g at time t.

[0100] T odd For odd-numbered moments,

[0101] T even The time is an even number.

[0102] A stochastic optimization scheduling system for determining wind power operating reserve capacity includes: a data acquisition module, a wind power operating reserve capacity calculation module, a key scenario set screening module, and a scheduling optimization module;

[0103] The data acquisition module is used to collect actual operating data and forecast data of wind turbine units;

[0104] The wind power operation reserve capacity calculation module is used to establish the error probability distribution function of the intraday wind power prediction power based on the actual operation data and prediction data of the wind turbine units, and to determine the credibility of the intraday wind power prediction power by using the inverse function of the error probability distribution function and the set prediction power accuracy probability; and to use the output of the wind turbine units that is greater than the credibility as the operation reserve capacity provided by the wind power.

[0105] The key scenario set filtering module is used to filter out key scenario sets from the intraday wind power output scenario set using a full-scenario feasible random optimization method;

[0106] The scheduling optimization module is used to establish a unit combination model for wind power in operation reserve for key scenario sets, with the objective function of minimizing the operating cost of the power system; during the intraday operation phase, the unit combination model for wind power in operation reserve is used to determine the operation reserve capacity of synchronous thermal power units and wind power units.

[0107] The beneficial effects of this invention are as follows: Compared with the prior art, this invention firstly obtains the probability density estimate of wind power prediction error based on data-driven methods, and incorporates wind power into the system's operational reserve at a certain confidence interval level; secondly, it establishes a stochastic unit combination model for systems with a high proportion of wind power, introduces strong unexpected conditions to transform the model, and uses a full-scenario feasible stochastic optimization method to solve the transformed model. The proposed full-scenario feasible stochastic optimization method for incorporating wind power output into the reserve system ensures the unexpectedness of economic dispatch decisions and the full-scenario feasibility of unit combination decisions. Attached Figure Description

[0108] Figure 1 This is a flowchart of a stochastic optimization scheduling method for determining wind power operation reserves proposed in this invention. Detailed Implementation

[0109] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of this invention. The embodiments described in this application are merely some embodiments of this invention, and not all embodiments. Based on the spirit of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of this invention.

[0110] This invention proposes a stochastic optimization scheduling method for determining wind power operating reserves, such as... Figure 1 As shown, it includes:

[0111] Step 1: Collect actual operating data and forecast data of wind turbine units, and establish the error probability distribution function of intraday wind power forecast based on kernel density estimation.

[0112] Kernel density estimation is a nonparametric estimation method that has a better fit than classical hypothetical models and higher accuracy in inferring past production activities.

[0113] The basic idea of ​​kernel density estimation is: if a sample is observed to exist in the sample set, then this sample can be considered to be the maximum point of the probability density. Numbers closer to the sample have a higher probability of appearing, while numbers farther away from the sample have a lower probability of appearing. This relationship can be expressed in the form of a normal distribution probability density function, which is called the kernel function.

[0114] Assume the cumulative distribution function of the sample data is The probability density function is Then the following relationship holds:

[0115]

[0116] In the formula, h is the bandwidth of the kernel density estimation, x is any sample, and n is the number of samples.

[0117] Once the sample set is determined, the kernel function K(·) and h jointly determine the quality of the kernel density estimate. From the definition of kernel density estimation, K(·) is generally chosen as a symmetric unimodal probability density function centered at 0, such as a uniform function or a Gaussian function. To ensure the continuity of the probability density, it possesses the following properties:

[0118] ∫K(x)dx=1

[0119] ∫xK(x)dx=0

[0120] ∫x 2 K(x)dx=c

[0121] In the formula, c is a constant greater than zero.

[0122] When h is the optimal bandwidth, the kernel density estimation can be guaranteed to have stable consistency for any sample, and choosing any kernel function K(·) has an equivalent effect. Therefore, this invention uses a Gaussian function as the kernel function, as follows:

[0123]

[0124] In the formula, x i Let i be the i-th sample.

[0125] Based on nonparametric statistical theory, the optimal bandwidth h is calculated by minimizing the asymptotic mean square integral error, as follows:

[0126]

[0127]

[0128] In the formula,

[0129] ASE(x) minimizes the asymptotic mean square integral error.

[0130] E(·) is the expected value function.

[0131] o(·) represents the remainder term in the integral, indicating (1 / nh+h) 4 () higher-order infinitesimals,

[0132] σ is the standard deviation.

[0133] R(·) describes the roughness of the function. Combining the above formula, the optimal bandwidth h for kernel density estimation can be obtained as:

[0134]

[0135] By analyzing the power prediction error, the accuracy of the predicted power can be indirectly reflected. Subsequently, the proportion of wind power that can be used as a reserve is determined by the probability distribution function of the predicted power error. This invention proposes a method to determine the probability distribution function of wind turbine power prediction error using kernel density estimation and to screen out a key set of wind power output scenarios, which reduces the difficulty of subsequent stochastic optimization and improves the model's computational efficiency.

[0136] Step 2: Using the inverse function of the error probability distribution function and the set prediction power accuracy probability, determine the credibility of the intraday wind power prediction power; use the output of wind turbine units that is greater than the credibility as the operating reserve capacity provided by wind power.

[0137] The preset reliability of wind power output is determined by the error probability distribution function of the predicted power in step 1. The preset reliability of wind power output can be represented by finding the inverse function of the probability distribution function F(x). The remaining portion of the operating reserve capacity is provided by thermal power units, thus determining the maximum proportion of wind power that can be included in the operating reserve capacity.

[0138] If we take A p The reliability of intraday wind power forecasts is expressed as A. Reliability refers to the accuracy of power forecasts with a certain probability C%. p The corresponding wind turbine output is taken as the reliable output and included in the operational reserve capacity. Selecting C% = 95%, A... p At this point, it can be calculated using the following formula:

[0139] A p =G(1-0.95)

[0140] In the formula, G(·) is the inverse function of the probability distribution function F(x) of wind power prediction error, and F(x) is the probability density function. The integral function.

[0141] At this time, the operating reserve capacity provided by wind power is expressed as:

[0142]

[0143] In the formula, w is the operating reserve capacity provided by wind turbine j at time t. j,t Let be the predicted power of wind turbine j at time t.

[0144] The remaining operating reserve is provided by thermal power units, and the operating reserve model of thermal power units is represented as follows:

[0145]

[0146]

[0147] In the formula

[0148] R ug,t R dg,t These represent the positive and negative operating reserve capacity of thermal power unit g at time t, respectively.

[0149] and These are the upper and lower limits of the output of thermal power unit g, respectively.

[0150] λ g Let g be the ramp rate of the thermal power unit.

[0151] P g,t Let g be the output of the thermal power unit g at time t.

[0152] t R This is for running a backup response time.

[0153] At this point, the system's operational standby capacity satisfies the following constraints:

[0154]

[0155]

[0156] In the formula,

[0157] G represents the number of thermal power units.

[0158] K w The number of wind turbine units.

[0159] These represent the positive and negative operating reserve capacities required by the system at time t, respectively.

[0160] Step 3: Using the full-scenario feasible stochastic optimization method, select key scenario sets from the intraday wind power output scenario set; for the key scenario sets, establish a unit combination model for wind power to be included in the operating reserve with the objective function of minimizing the operating cost of the power system; during the intraday operation phase, use the unit combination model for wind power to be included in the operating reserve to determine the operating reserve capacity of synchronous thermal power units and wind power units.

[0161] The key scenario set is generated using the All-scenario-feasible Stochastic Optimization (ASFSO) method. The key to ASFSO lies in the selection and construction of key scenarios. ASFSO is an uncertain optimization method that combines the advantages of scenario-based stochastic optimization and robust optimization methods. It overcomes the drawbacks of stochastic optimization, such as a large number of scenarios and large computational scale, and robust optimization, where the uncertain set may not meet unpredictable conditions. The core idea of ​​ASFSO is to select appropriate vertices based on the vertex characteristics of the uncertain set of random variables, establish a finite number of scenarios, and use the minimum expected cost or maximum expected benefit as the objective function to obtain the optimal decision for the current period, ensuring its feasibility under all possible values ​​of future random factors. This invention modifies the two-stage stochastic optimization model into the ASFSO method, which improves upon the shortcomings of robust optimization / scenario-based stochastic optimization methods. The combination of this method with wind power backup optimization is a relatively unique research topic.

[0162] Specifically, step 3 includes:

[0163] Step 3.1: Using wind power output as a random variable, a convex polyhedral uncertainty set U is used to describe the intraday wind power output scenario set. Using a full-scenario feasible stochastic optimization method, the key scenario set selected from the intraday wind power output scenario set includes: basic scenario S. BS Vertex Scene S SVS and extreme climbing scenario S ERS The key scenario set is represented as S = {S BS ,S SVS ,S ERS}

[0164] Step 3.1.1: Determine the basic scenario S based on historical operational data. BS .

[0165] S BS This refers to scenarios with special meanings or specific scenarios, including but not limited to the expected value of predicted wind power output or scenarios generated from historical wind power operation data. Generally, scenarios that occur frequently and are typical are selected as basic scenarios based on historical operation data. During the random optimization and scenario generation process, only one basic scenario, i.e., the expected scenario, is considered, as follows:

[0166]

[0167] In the formula,

[0168] w BS This is a matrix representation of the wind turbine output in the basic scenario.

[0169] Let j be the output of wind turbine j at time t in the basic scenario.

[0170] K w The number of wind turbine units.

[0171] T represents the number of time periods.

[0172] E(·) is the expected value function.

[0173] w j,t The predicted output of wind turbine j at time t.

[0174] Step 3.1.2: Select vertices at the same sorting order m at each time step from the uncertain set U of the convex polyhedron to form the vertex scene S. SVS .

[0175] S SVS It is selected from the uncertain set U of the convex polyhedron according to certain rules. Assume the uncertain set U of the convex polyhedron at time t... t There is N t vertices For all vertices V at time t n,t According to the sum of the elements Arrange from smallest to largest. In the formula, Let represent the output of wind turbine j at time t under different vertices.

[0176] The vertices at the same sorting position m at each time point constitute the scene of each vertex, which can be represented as follows:

[0177]

[0178] In the formula, in the formula,

[0179] w SVS This represents the matrix representation of the wind turbine output in the vertex scenario.

[0180] Let m be the output of wind turbine j at time t in the vertex scenario, and let m be the sorting order.

[0181] N t Let U be the uncertainty set of a convex polyhedron at time t. t The number of vertices in the array.

[0182] Step 3.1.3 defines an odd-numbered scenario as one where wind power output has an upper limit during odd-numbered periods and a lower limit during even-numbered periods, and an even-numbered scenario as one where wind power output has a lower limit during odd-numbered periods and an upper limit during even-numbered periods. The odd-numbered and even-numbered scenarios together constitute the extreme ramping scenario S. ERS .

[0183] SERS Two scenarios of extreme fluctuations in wind power output are described, which are important conditions for meeting the unpredictability requirement. ERS The generation process is represented as follows:

[0184]

[0185]

[0186]

[0187] In the formula,

[0188] w OS This represents the matrix representation of wind turbine output in odd-number scenarios.

[0189] Let represent the output of wind turbine j at time t in the odd-number scenario.

[0190] w ES This represents the matrix representation of wind turbine output in even-number scenarios.

[0191] Let be the output of wind turbine j at time t in even-numbered scenarios.

[0192] T odd For odd-numbered moments,

[0193] T even For even-numbered moments,

[0194] These represent the upper and lower limits of the output of wind turbine j at time t, respectively.

[0195] Using the above scenario generation method, a total of feasible random optimization methods for all scenarios were generated. A set of key scenarios, including: There are 1 vertex scene, 1 basic scene, and 2 extreme climbing scenes; among them... Compared with the existing technology of scene-based random optimization, the present invention significantly reduces the number of scenes, effectively reducing the computational scale.

[0196] Step 3.2: For the key scenario set, with the objective function of minimizing the operating cost of the power system, establish a unit combination model for wind power to be included in the operating reserve. The unit combination model for wind power to be included in the operating reserve includes: a dispatch model and the constraints of the dispatch model.

[0197] Step 3.2.1, for the key scenario set, the scheduling model is as follows:

[0198]

[0199] In the formula,

[0200] C state For unit start-up and shutdown costs,

[0201] C cur For the cost of wind power abandonment,

[0202] C oper For coal consumption costs of thermal power units,

[0203] C back For system backup costs,

[0204] π s Let be the probability of the key scenario s.

[0205] The costs satisfy the following relationship:

[0206]

[0207] In the formula,

[0208] a, b, and c are the coefficients of the quadratic, linear, and constant terms of coal consumption cost, respectively.

[0209] P g,t Let g be the output of the thermal power unit g at time t.

[0210] These are the start-up cost and shutdown cost of thermal power unit g, respectively.

[0211] Let be the cost of wind power curtailment per unit of electricity at time t.

[0212] ΔR j,t Let be the amount of wind power curtailed by wind turbine j at time t.

[0213] c ug c dg These represent the purchase cost of the positive operating reserve capacity and the purchase cost of the negative operating reserve capacity of thermal power unit g, respectively.

[0214] R ug,t R dg,t These represent the positive and negative operating reserve capacity of thermal power unit g at time t, respectively.

[0215] c wj The cost of purchasing the operating reserve capacity of wind turbine unit j.

[0216] The operating reserve capacity provided for wind turbine j at time t.

[0217] G represents the number of thermal power units.

[0218] K wThe number of wind turbine units.

[0219] T represents the number of time periods.

[0220] Step 3.2.2, the constraints of the scheduling model include:

[0221] 1. The unit operating state logic constraint characterizes the operating state x of the thermal power unit g in the system. g,t With startup status u g,t and closed state v g,t Constraints between them;

[0222]

[0223]

[0224] In the formula, x represents the temperature at which the thermal power unit g starts up. g,t =1, x when shutting down g,t =0; where, G represents the number of thermal power units.

[0225] 2. Constraints on minimum start-up and shutdown times of thermal power units:

[0226]

[0227]

[0228] In the formula,

[0229] These are the minimum start-up and shutdown times for thermal power unit g, respectively.

[0230] The above two types of constraints represent all constraints that are only related to the start-up variables of thermal power units. Since the start-up and shutdown plans of thermal power units are the same for any wind power output scenario, they can be regarded as deterministic constraints here.

[0231] 3. Constraints on the output of thermal power units:

[0232]

[0233] In the formula, Let g be the minimum output of the thermal power unit. P is the maximum output of the thermal power unit g. g,t Let g be the output of the thermal power unit g at time t.

[0234] 4. Constraints on wind turbine output:

[0235]

[0236] In the formula, Let be the power of wind turbine j at time t. Let be the predicted power of wind turbine j at time t.

[0237] 5. Constraints on the climbing slope of thermal power units:

[0238]

[0239] In the formula,

[0240] These represent the upward and downward ramp rates of unit power during the initial and final periods and the middle period of operation of thermal power unit g, respectively.

[0241] 6. System backup constraints:

[0242]

[0243]

[0244] Among them, constraints 3 to 6 are full-scenario feasible constraints, which require that they be satisfied for any key scenario s∈U of wind power output.

[0245] 7. Strongly unexpected constraints:

[0246]

[0247]

[0248]

[0249] In the formula,

[0250] These represent the odd-numbered scenarios S respectively. OS Even-numbered scenarios S ES The output of the thermal power unit g at time t.

[0251] x g,t Let g be the operating state of the thermal power unit at time t, and x be the operating state at startup. g,t =1, x when shutting down g,t =0,

[0252] S BS S SVS S ERS These are the basic scene, the vertex scene, and the extreme climbing scene.

[0253] P g,t Let g be the output of the thermal power unit g at time t.

[0254] T odd For odd-numbered moments,

[0255] T even The time is an even number.

[0256] The essence of introducing strong unpredictability conditions to transform the model is to decouple the ramping constraints between time periods, ensuring that scheduling decisions meet the unpredictability. Furthermore, unlike the min-max-min structure in robust optimization, the transformed second-stage economic scheduling problem can be viewed as a single-time-period problem, eliminating the need to consider in advance the future unit output values ​​to satisfy the ramping constraints.

[0257] Step 3.3: During the intraday operation phase, the operating reserve capacity of synchronous thermal power units and wind power units is determined using the unit combination model that includes wind power in the operating reserve.

[0258] This invention addresses the uncertainty issues arising from large-scale wind power grid connection through a full-scenario feasible stochastic optimization method. Specifically, during the intraday operation phase, it calculates the operating reserve capacity of synchronous thermal power units and wind power units to more accurately formulate operating reserve strategies and improve wind power absorption rate.

[0259] This disclosure can be a system, method, and / or computer program product. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of this disclosure.

[0260] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example—but not limited to—electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination of the foregoing. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0261] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0262] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.

[0263] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A stochastic optimization scheduling method for determining wind power operating reserves, characterized in that, include: Step 1: Collect actual operating data and forecast data of wind turbine units, and establish the error probability distribution function of intraday wind power forecast based on kernel density estimation; Step 2: Determine the reliability of the intraday wind power forecast using the inverse function of the error probability distribution function and the set prediction power accuracy probability; use the output of wind turbine units that are greater than the preset reliability as the operating reserve capacity provided by wind power. Step 3: Using the full-scenario feasible stochastic optimization method, select the key scenario set from the intraday wind power output scenario set; for the key scenario set, establish a unit combination model for wind power to be included in the operation reserve with the objective function of minimizing the operating cost of the power system. During the daytime operation phase, the operating reserve capacity of synchronous thermal power units and wind power units is determined using a unit combination model that incorporates wind power into the operating reserve. Among them, using a full-scenario feasible stochastic optimization method, the key scenario set selected from the intraday wind power output scenario set includes: Determine the basic scenario based on historical operational data. In the process of randomly optimizing and generating scenarios, only one basic scenario is considered, which is the desired scenario, as follows: ， In the formula, This is a matrix representation of the wind turbine output in the basic scenario. Wind turbine units in basic scenarios At any moment of efforts, The number of wind turbine units. Number of time periods It is the expected value function. For wind turbines At any moment The predicted output; From the uncertain set of convex polyhedra Selecting each time point in the same sorting order The vertices constitute the vertex scene. ,as follows: In the formula, This represents the matrix representation of the wind turbine output in the vertex scenario. For wind turbines in the vertex scenario At any moment The output, and the sorting order is , For a moment Uncertainty set of convex polyhedra The number of vertices in the array; Scenarios where wind power output has an upper limit during odd-numbered periods and a lower limit during even-numbered periods are defined as odd-numbered scenarios, while scenarios where wind power output has a lower limit during odd-numbered periods and an upper limit during even-numbered periods are defined as even-numbered scenarios. These odd-numbered and even-numbered scenarios together constitute the extreme ramping scenario. ; The generation process is represented as follows: ， ， ， In the formula, This represents the matrix representation of wind turbine output in odd-number scenarios. Wind turbines in odd-number scenarios At any moment of efforts, This represents the matrix representation of wind turbine output in even-number scenarios. Wind turbines in even-number scenarios At any moment of efforts, For odd-numbered moments, For even-numbered moments, , Wind turbine At any moment The upper and lower limits of output.

2. The stochastic optimization scheduling method for determining wind power operating reserves according to claim 1, characterized in that, Using a Gaussian function as the kernel function, the optimal bandwidth for kernel density estimation is: In the formula, For the bandwidth of kernel density estimation, For the sample size, The standard deviation is denoted as .

3. The stochastic optimization scheduling method for determining wind power operating reserves according to claim 1, characterized in that, The reliability of intraday wind power forecasts is as follows: In the formula, To assess the reliability of intraday wind power forecasts, This is the inverse function of the probability distribution function of wind power prediction error. The set probability of predicted power accuracy.

4. The stochastic optimization scheduling method for determining wind power operating reserves according to claim 3, characterized in that, The probability of predicting power accuracy is set at 95%.

5. The stochastic optimization scheduling method for determining wind power operating reserves according to claim 3, characterized in that, The constraints on the operating reserve capacity provided by wind power are as follows: In the formula, For wind turbines At any moment Provided operational standby capacity, For wind turbines At any moment The predicted power.

6. The stochastic optimization scheduling method for determining wind power operating reserves according to claim 1, characterized in that, Step 3 includes: Step 3.1: Using the wind turbine output as a random variable, employ the convex polyhedron uncertainty set... Describe the intraday wind power output scenario set; using a full-scenario feasible stochastic optimization method, the key scenario set selected from the intraday wind power output scenario set includes: basic scenario Vertex Scene and extreme climbing scenarios The key scenario set is represented as ; Step 3.2: For the key scenario set, with the objective function of minimizing the operating cost of the power system, establish a unit combination model for wind power to be included in the operating reserve; the unit combination model for wind power to be included in the operating reserve includes: a dispatch model and the constraints of the dispatch model; Step 3.3: During the intraday operation phase, the operating reserve capacity of synchronous thermal power units and wind power units is determined using the unit combination model that includes wind power in the operating reserve.

7. The stochastic optimization scheduling method for determining wind power operating reserves according to claim 1, characterized in that, Generate using a feasible stochastic optimization method across all scenarios Three key scenario sets, including: One vertex scene, one basic scene, and two extreme climbing scenes; in, .

8. The stochastic optimization scheduling method for determining wind power operating reserves according to claim 6, characterized in that, The scheduling model is as follows: In the formula, For unit start-up and shutdown costs, For the cost of wind power abandonment, For coal consumption costs of thermal power units, For system backup costs, For key scenarios The probability of.

9. The stochastic optimization scheduling method for determining wind power operating reserves according to claim 8, characterized in that, The costs satisfy the following relationship: In the formula, , , These are the coefficients for the quadratic term, the linear term, and the constant term of the coal consumption cost, respectively. For thermal power units At any moment of efforts, , thermal power units Start-up and downtime costs, For a moment The cost of wind power curtailment per unit of electricity, For wind turbines At any moment Wind power curtailment , thermal power units The purchase cost of positive operating reserve capacity and the purchase cost of negative operating reserve capacity. , thermal power units At any moment Positive operating reserve capacity, negative operating reserve capacity, For wind turbines The cost of purchasing operating standby capacity. For wind turbines At any moment Provided operational standby capacity, This refers to the number of thermal power units. The number of wind turbine units. This represents the number of time periods.

10. The stochastic optimization scheduling method for determining wind power operating reserves according to claim 8, characterized in that, The constraints of the scheduling model include: 1) Constraints between the operating state, start-up state, and shutdown state of thermal power units; 2) Constraints on the minimum start-up and shutdown times of thermal power units: 3) Constraints on the output of thermal power units; 4) Constraints on wind turbine output; 5) Constraints on the climbing ability of thermal power units; 6) Constraints on system standby capacity.

11. The stochastic optimization scheduling method for determining wind power operating reserves according to claim 10, characterized in that, The constraints of the scheduling model also include: strongly unexpected constraints; as follows: In the formula, , These represent odd-numbered scenarios. Even-numbered scenes Lower thermal power unit At any moment of efforts, For thermal power units At any moment The operating status, when booting up =1, when the device is off =0, , , These are the basic scene, the vertex scene, and the extreme climbing scene. For thermal power units At any moment of efforts, For odd-numbered moments, The time is an even number.

12. A stochastic optimization scheduling system for determining wind power operating reserve using the method described in any one of claims 1-11, characterized in that, include: Data acquisition module, wind power operation reserve capacity calculation module, key scenario set filtering module, and scheduling optimization module; The data acquisition module is used to collect actual operating data and forecast data of wind turbine units; The wind power operation reserve capacity calculation module is used to establish an error probability distribution function of the intraday wind power prediction power based on the actual operation data and prediction data of the wind turbine, and to determine the credibility of the intraday wind power prediction power by using the inverse function of the error probability distribution function and the set prediction power accuracy probability; and to use the output of the wind turbine that is greater than the preset credibility as the operation reserve capacity provided by the wind power. The key scenario set filtering module is used to filter out key scenario sets from the intraday wind power output scenario set using a full-scenario feasible random optimization method; The scheduling optimization module is used to establish a unit combination model for wind power to be included in the operating reserve for key scenario sets, with the objective function of minimizing the operating cost of the power system. During the daytime operation phase, the operating reserve capacity of synchronous thermal power units and wind power units is determined using a unit combination model that incorporates wind power into operational reserves.

Images