Wind power system low-carbon optimization scheduling method considering demand response

Abstract

The invention discloses a wind power system low-carbon optimization scheduling method considering demand response, and the method comprises the following steps: collecting the prediction data and prediction deviation information of the day-ahead wind power output and load of a wind power system, and building an uncertainty model of a source load; performing random sampling on the wind power and the load by using a Latin hypercube sampling method to obtain an initial scene set of the wind power and the load; reducing the initial scene set of the wind power and the load by adopting a heuristic synchronous back substitution method, generating typical scenes of the wind power and the load, and simultaneously obtaining occurrence probability information of each typical scene; establishing a demand side respondable load model; based on the established demand side responsible load model, establishing an improved stepped demand response excitation mechanism model; constructing a two-stage optimization model by adopting a multi-scene stochastic optimization method, wherein the two-stage optimization model comprises a demand response upper-layer optimization model and a power system lower-layer optimization model; and solving the two-stage optimization model by adopting a Cplex solver, and outputting an optimization scheduling result of the wind power system. Through demand response and multi-element energy coordinated optimization, the supply and demand curve matching degree of the system is improved, and the wind power consumption rate and the low-carbon and economic performance of operation of the system are effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power system optimization and dispatching, and in particular to a low-carbon optimization and dispatching method for a wind power system taking demand response into consideration. Background Art

[0002] The volatility and intermittent characteristics of renewable energy power generation are becoming increasingly prominent, and the degree of fit with the power load curve has significantly decreased, which has intensified the pressure on the balance of power supply and demand in the power system and posed new challenges to the effective absorption and efficient utilization of renewable energy. In view of this, for wind power systems, incentive mechanisms are used to guide user groups to actively participate in demand response plans, so as to flexibly adjust the load structure to better adapt to the unstable characteristics of renewable energy power generation. At the same time, the operation strategies of traditional units, new energy units and energy storage power stations are optimized to achieve coordinated optimization of the output among the three, aiming to effectively promote the large-scale absorption of wind power, improve the low-carbon economic performance of the entire power system operation, and achieve a win-win situation of environmental and economic benefits.

[0003] In the current research field of power system optimization and dispatch, demand response, as a promising means for power systems containing new energy units, can realize the auxiliary consumption of new energy by flexibly adjusting the flexible load in the load structure. However, demand response faces the problem of imperfect incentive mechanism in application, resulting in low willingness and enthusiasm of users to participate. In order to better promote the further development of demand response, it is necessary to improve the demand response incentive mechanism, increase user participation, and broaden the adjustable range of flexible resources on the demand side. The low-carbon economic dispatch strategy that considers user demand response and coordinated optimization of multiple energy sources is of great significance for coping with the challenges of volatility of new energy power generation, promoting efficient use of energy, and reducing carbon emissions and operating costs. Summary of the invention

[0004] Purpose of the invention: The present invention provides a low-carbon optimization scheduling method for wind power systems taking demand response into consideration. By means of demand response and coordinated optimization of multiple energy sources, the matching degree of the supply and demand curves of the system is improved, and the wind power consumption rate of the system and the low-carbon and economic performance of its operation are effectively improved.

[0005] Technical solution: A low-carbon optimization scheduling method for a wind power system considering demand response described in the present invention comprises the following steps:

[0006] Step 1: Collect the forecast data and forecast deviation information of the wind power output and load of the wind power system, and establish the uncertainty model of source and load, including the uncertainty model of wind power output and the uncertainty model of load demand;

[0007] Step 2: Use Latin hypercube sampling method to randomly sample wind power and load respectively to obtain initial scenario sets of wind power and load;

[0008] Step 3: Use the heuristic synchronous back-substitution method to reduce the initial scenario set of wind power and load, generate typical scenarios of wind power and load, and obtain the occurrence probability information of each typical scenario;

[0009] Step 4: Establish a demand-side responsive load model, wherein the demand-side responsive load model includes a curtailable load model and a transferable load model;

[0010] Step 5: Based on the established demand-side responsive load model, an improved step-type demand response incentive mechanism model is established;

[0011] Step 6: Use a multi-scenario stochastic optimization method to build a two-stage optimization model, including a demand response upper-level optimization model and a power system lower-level optimization model;

[0012] Step 7: Use the Cplex solver to solve the two-stage optimization model and output the optimization scheduling result of the wind power system.

[0013] Furthermore, in step 1, the uncertainty model of the source load is established as follows:

[0014] (a) Uncertainty model of wind power output, which is expressed as follows:

[0015]

[0016] In the formula, They represent the actual output value and predicted output value of wind power at time t respectively; represents the wind power output prediction error; σ w (t) represents the standard deviation of wind power output prediction error; represents the total installed capacity of the wind farm;

[0017] (b) The uncertainty model of load demand is expressed as follows:

[0018]

[0019] In the formula, and They represent the actual value and predicted value of the system load during period t respectively; represents the load forecast error; σ e (t) represents the standard deviation of load forecast error; e L It is expressed as a proportionality factor that characterizes the accuracy of the prediction tool.

[0020] Furthermore, in step 2, the Latin hypercube sampling method is used to randomly sample wind power and load, respectively, and the initial scene sets of wind power and load are obtained, which specifically include the following steps:

[0021] Step 21, set the power to follow the normal distribution N~(μ,δ), where μ is the expected value of the predicted value, δ is the fluctuation percentage, the number of samples is set to 1000, and the vertical axis of the cumulative probability distribution curve is divided into 1000 intervals with equal probability, and the width of each interval is 1 / 1000;

[0022] Step 22: randomly select a value in each interval of the variable one by one, and select 1000 values ​​in total;

[0023] Step 23: The cumulative distribution probability corresponding to the kth sampling value is F i (k)=(r u +k-1) / 1000, k=1,2,…,1000, where: r u ~N(0,1) obeys uniform distribution, and the kth sampling value of the random variable can be obtained through the inverse function;

[0024] Step 24: After sampling, the sampling values ​​of all random variables are formed into a sampling matrix, and the sampling values ​​of each column of the matrix are sorted by using the Gram-Schemidt sequence orthogonalization method, and the correlation of each column is minimized through iterative calculation;

[0025] Step 25: Form an initial scene set containing 1000 randomly sampled scenes.

[0026] Furthermore, in step 3, the initial scenario set of wind power and load is reduced by using the heuristic synchronous back-substitution method to generate typical scenarios of wind power and load, and the occurrence probability information of each typical scenario is obtained, including the following steps:

[0027] Step 31, define the number of target scenes to be reduced as 10, and the scene probability distance as the binary norm of the scene vector;

[0028] Step 32: Calculate the Euclidean distance between any two scenes

[0029] Step 33: Select the scene S i The closest scene S s , that is, D is =minD ij , calculate the probability distance p between scenes i and s is =D is p s , where D is It is scene S i and S s The Euclidean distance between s It is scene S s The probability of occurrence;

[0030] Step 34: Find the probability distance value p in the sample isThe smallest scene pair (S i ,S s );

[0031] Step 35: Let p s =p s +p i , in the update p s At the same time cut to scene S i , and correct the number of samples in the reduction process;

[0032] Step 36, return to step 32 to perform iterative calculation, and stop the iteration when the number of scenes to be reduced is reached.

[0033] Furthermore, in step 4, the demand-side responsive load model is established as follows:

[0034] (a) The curtailable load model is expressed as follows:

[0035]

[0036] In the formula, ξ a (t), P a (t) They represent the load reduction state, load reduction amount, minimum load reduction amount, maximum load reduction amount and total load reduction amount at time t respectively;

[0037] (b) Transferable load model, which is expressed as follows:

[0038]

[0039] In the formula, ξ bin (t) and ξ bout (t) are all 0-1 state variables, ξ bin (t) = 1 indicates load transfer, ξ bout (t) = 1 indicates load transfer; P bin (t), P bout (t) are the load transfer in and out at time t; P b (t) is the total load transfer at time t; are the upper and lower limits of the load transferred in at time t; is the upper and lower limits of the output load at time t; N T Indicates the scheduling period;

[0040] (c) The demand side responsive load model is expressed as follows:

[0041]

[0042] In the formula, α represents the percentage of the total amount of defaulted and unresponsive load that cannot perform full response after the user signs the demand response contract to the total amount of signed response load; It indicates the total amount of flexible load actually provided by users in the demand response process during the dispatch period.

[0043] Furthermore, in step 5, the improved step-type demand response incentive mechanism model is established as follows:

[0044] (a) The reward mechanism model is expressed as follows:

[0045]

[0046] In the formula, f DR is the economic compensation for users participating in demand response; ε is the base price of compensation per unit of response electricity; v is the growth rate of the compensation unit price; d is the length of the incentive interval under the reward mechanism;

[0047] (b) The penalty mechanism model is expressed as follows:

[0048]

[0049] In the formula, is the economic penalty for the user's failure to respond to the load, and its value is a negative number; τ represents the penalty base price per unit of default electricity; D represents the length of the step penalty interval; It represents the total load that is increased due to the user not fulfilling the response condition but going in the opposite direction.

[0050] Furthermore, in step 6, the objective function of the upper optimization model of demand response is expressed as follows:

[0051]

[0052] In the formula, f1 represents the response benefit of users on the demand side; Represents the discomfort cost caused by users adjusting their electricity consumption behavior; ω a ,ω b are the unit electricity discomfort coefficients for load shedding and shifting, respectively;

[0053] The objective function of the lower-level optimization model of the power system is expressed as follows:

[0054]

[0055] Where f2 represents the comprehensive operating cost of the system; ρ j represents the distribution probability of the jth typical scene; N j The total number of scenes representing typical scenes; f j,g represents the power generation cost of traditional thermal power units in the jth typical scenario; fj,ms represents the operating cost of the hybrid energy storage system; represents the carbon emission penalty cost generated by the traditional thermal power unit in the power generation process under the jth typical scenario; w b 、w s are the operation and maintenance cost coefficients of the electrochemical energy storage power station and the pumped storage power station; P b (t), P s (t) is the output power of the electrochemical energy storage power station and the pumped storage power station at time t; is the carbon emission penalty cost coefficient; is the carbon emission coefficient per unit of electricity of the i-th thermal power unit; It represents the total carbon emissions of all thermal power units in the system at time t.

[0056] Furthermore, the constraints of the lower-level optimization model of the power system include:

[0057] A. Wind turbine operation constraints, the expressions are as follows:

[0058]

[0059] In the formula, is the upper limit of wind power output;

[0060] B. Traditional unit operation constraints, including unit output upper and lower limit constraints, unit climbing constraints and minimum start and stop time constraints, are expressed as follows:

[0061]

[0062] In the formula, u g,i (t) and u g,i (t-1) represents the start and stop status of the i-th traditional unit at time t and time (t-1), respectively; They represent the minimum output power and maximum output power of the i-th conventional unit respectively; is the maximum climbing power of the traditional unit; t on and t off is the minimum continuous duration of unit startup and shutdown; t0 is the current time t;

[0063] C. Energy storage operation constraints, including electrochemical energy storage power station and pumped storage power station operation constraints, are expressed as follows:

[0064]

[0065] 0≤u b,dis (t)+u b,ch (t)≤1

[0066]

[0067] S b (1) = S b (N T )

[0068]

[0069] 0≤u vp (t)+u vg (t)≤1

[0070] V s min ≤V s (t)≤V s max

[0071] V s (1) = V s (N T )

[0072] In the formula, u b,ch (t),u b,dis (t) respectively represent the charging and discharging states of the electrochemical energy storage power station at time t; and They are the maximum charging power and maximum discharging power of the electrochemical energy storage power station respectively; and u is the upper and lower limits of the storage capacity of the electrochemical energy storage power station; vp (t),u vg (t) respectively represent the pumping and power generation status of the pumped storage power station at time t; P vp (t) and P vg (t) represent the pumping and generating power of the pumped storage power station at time t, respectively; They represent the upper and lower limits of the pumping power of the pumped storage power station respectively; Respectively represent the upper and lower limits of the pumping and power generation of the pumped storage power station; V s max With V s min The upper and lower limits of the water storage capacity of the pumped storage power station;

[0073] D. Power balance constraint, its expression is as follows:

[0074]

[0075] E. Line flow and transmission capacity constraints, the expressions are as follows:

[0076]

[0077] Where: P iInject power into power system nodes; B ij 、x ij , P ij They are the branch susceptance, branch reactance and branch transmission power of the power system, among which, B ij =-1 / x ij , θ i ,θ j , Δθ ij They are respectively the phase angle of the node at the beginning of the branch, the phase angle of the node at the end of the branch, and the difference between the phase angles at the beginning and end of the branch.

[0078] Furthermore, in step 7, the Cplex optimization solver is used to solve the demand response upper-level optimization model, and the optimized load information is output, including the reducible load response amount, the transferable load response amount, the optimized load curve, and the demand response compensation cost information.

[0079] Furthermore, the Cplex solver is used to solve the lower-level optimization model of the power system and output the optimization scheduling result information of the wind power system, including the output plan of wind turbine units, the output plan of traditional units, the output plan of electrochemical energy storage power station, the output plan of pumped storage power station, various energy consumption of the system, carbon emissions, and economic operation cost information.

[0080] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages: (1) In the power system optimization and scheduling method provided by the present invention, the coordinated optimization scheduling between demand response and multi-form energy storage, traditional units, and wind turbines under the improved step-type demand response incentive mechanism is considered, which is more superior in optimizing load structure, promoting system wind consumption, reducing carbon emissions, and improving the low-carbon economy of system operation; (2) In the power system optimization and scheduling method provided by the present invention, the user's demand response enthusiasm is greatly improved and the user's default probability is reduced; (3) In the power system optimization and scheduling method provided by the present invention, the demand response compensation cost can be reduced under the condition of ensuring a high response rate of user demand response, thereby greatly improving the economy of system operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0081] Figure 1 It is a schematic diagram of the method flow of the present invention.

[0082] Figure 2 This is a typical scenario of wind power output according to the present invention.

[0083] Figure 3 This is a typical scenario of the load demand of the present invention.

[0084] Figure 4 This is a load curve comparison diagram before and after the user participates in the improved step-type demand response of the present invention.

[0085] Figure 5 This is a schematic diagram of load optimization results under the improved stepped demand response incentive mechanism of the present invention.

[0086] Figure 6 This is a schematic diagram of the optimized scheduling results of the system under the improved stepped demand response incentive mechanism of the present invention.

[0087] Figure 7 This is a schematic diagram of the optimized dispatching results of the energy storage power station under the improved stepped demand response incentive mechanism of the present invention. DETAILED DESCRIPTION

[0088] like Figure 1 As shown, a low-carbon optimization scheduling method for a wind power system considering demand response includes the following steps:

[0089] Step 1: Collect the forecast data and forecast deviation information of the wind power output and load of the wind power system on the day before, and establish a source-load uncertainty model;

[0090] A. Uncertainty Model of Wind Power Output

[0091]

[0092] In the formula, They represent the actual output value and predicted output value of wind power at time t respectively; represents the wind power output prediction error; σ w (t) represents the standard deviation of wind power output prediction error; Represents the total installed capacity of the wind farm.

[0093] B. Uncertainty Model of Load Demand

[0094]

[0095] In the formula, and They represent the actual value and predicted value of the system load during period t respectively; represents the load forecast error; σ e (t) represents the standard deviation of load forecast error; e L It is expressed as a proportionality factor that characterizes the accuracy of the prediction tool;

[0096] Step 2: Use Latin hypercube sampling method to randomly sample wind power and load respectively to obtain initial scenario sets of wind power and load;

[0097] A. Assume that the power follows a normal distribution N~(μ,δ), where μ is the expected value of the predicted value, δ is the fluctuation percentage, the number of samples is set to 1000, and the vertical axis of the cumulative probability distribution curve is divided into 1000 intervals with equal probability, and the width of each interval is 1 / 1000;

[0098] B. Randomly select a value in each interval of the variable one by one, and select 1000 values ​​in total;

[0099] C. The cumulative distribution probability corresponding to the kth sampling value is F i (k)=(r u +k-1) / 1000, k=1,2,…,1000. Among them: r u ~N(0,1) obeys uniform distribution, and the kth sampling value of the random variable can be obtained through the inverse function;

[0100] D. After sampling, the sampling values ​​of all random variables are formed into a sampling matrix, and the sampling values ​​of each column of the matrix are sorted using the Gram-Schemidt sequence orthogonalization method, and the correlation of each column is minimized through iterative calculation;

[0101] E. Form an initial scene set containing 1000 randomly sampled scenes.

[0102] Step 3: Use the heuristic synchronous back-substitution method to reduce the initial scenario set of wind power and load, generate typical scenarios of wind power and load, and obtain the occurrence probability information of each typical scenario;

[0103] A. Define the number of target scenes to be reduced as 10, and the scene probability distance as the two-norm of the scene vector;

[0104] B. Calculate the Euclidean distance between any two scenes

[0105] C. Selection and Scenario S i The closest scene S s , that is, D is =minD ij . Calculate the probability distance p between scenes i and s is =D is p s Among them, D is It is scene S i and S s The Euclidean distance between s It is scene S s The probability of occurrence;

[0106] D. Find the probability distance value p in the sample is The smallest scene pair (S i ,S s );

[0107] E. Let p s =p s +p i , in the update p s At the same time cut to scene S i , and correct the number of samples in the reduction process;

[0108] F. Return to step (b) for iterative calculation. When the number of scenes to be reduced is reached, stop the iteration.

[0109] Step 4: Establish a demand-side responsive load model;

[0110] A. Load reduction model

[0111]

[0112] In the formula, ξ a (t), P a (t) They represent the load reduction state, load reduction amount, minimum load reduction amount, maximum load reduction amount and total load reduction amount at time t respectively.

[0113] B. Transferable Load Model

[0114]

[0115] In the formula, ξ bin (t) and ξ bout (t) are all 0-1 state variables, ξ bin (t) = 1 indicates load transfer, ξ bout (t) = 1 indicates load transfer; P bin (t), P bout (t) are the load transfer in and out at time t; P b (t) is the total load transfer at time t; are the upper and lower limits of the load transferred in at time t; is the upper and lower limits of the output load at time t; N T Indicates the scheduling period.

[0116] C. Demand-side responsive load model

[0117]

[0118] In the formula, α represents the percentage of the total amount of defaulted and unresponsive load that cannot perform full response after the user signs the demand response contract to the total amount of signed response load; It indicates the total amount of flexible load actually provided by users in the demand response process during the dispatch period.

[0119] Step 5: Establish an improved step-type demand response incentive mechanism model;

[0120] A. Reward Mechanism Model

[0121]

[0122] In the formula, f DR is the economic compensation for users participating in demand response; ε is the base price for compensation per unit of response electricity; v is the growth rate of the compensation unit price; d is the length of the incentive interval under the reward mechanism.

[0123] B. Punishment Mechanism Model

[0124]

[0125] In the formula, is the economic penalty for the user's failure to respond to the load, and its value is a negative number; τ represents the penalty base price per unit of default electricity; D represents the length of the step penalty interval; It represents the total load that is increased due to the user not fulfilling the response condition but going in the opposite direction.

[0126] Step 6: Use multi-scenario random optimization method to build a two-stage optimization model;

[0127] A. Demand response upper optimization model

[0128] The objective function of the upper optimization model of demand response is expressed as follows:

[0129]

[0130] In the formula, f1 represents the response benefit of users on the demand side; Represents the discomfort cost caused by users adjusting their electricity consumption behavior; ω a ,ω b are the unit electricity discomfort coefficients of load reduction and transfer, respectively.

[0131] B. Power System Lower Level Optimization Model

[0132] B1. Objective function of the power system lower-level optimization model

[0133]

[0134]

[0135] Where f2 represents the comprehensive operating cost of the system; ρ j represents the distribution probability of the jth typical scene; N j The total number of scenes representing typical scenes; f j,g represents the power generation cost of traditional thermal power units in the jth typical scenario; fj,ms represents the operating cost of the hybrid energy storage system; represents the carbon emission penalty cost generated by the traditional thermal power unit in the power generation process under the jth typical scenario; w b 、w s are the operation and maintenance cost coefficients of the electrochemical energy storage power station and the pumped storage power station; P b (t), P s (t) is the output power of the electrochemical energy storage power station and the pumped storage power station at time t; is the carbon emission penalty cost coefficient; is the carbon emission coefficient per unit of electricity of the i-th thermal power unit; It represents the total carbon emissions of all thermal power units in the system at time t.

[0136] B2. Constraints of the Lower-Level Optimization Model of the Power System

[0137] (1) Wind turbine operation constraints, which are expressed as follows:

[0138]

[0139] In the formula, It is the upper limit of wind power output.

[0140] (2) Traditional unit operation constraints include unit output upper and lower limit constraints, unit ramp constraints, and minimum start and stop time constraints, which are expressed as follows:

[0141]

[0142] In the formula, u g,i (t) and u g,i (t-1) represents the start and stop status of the i-th traditional unit at time t and time (t-1), respectively; They represent the minimum output power and maximum output power of the i-th conventional unit respectively; is the maximum climbing power of the traditional unit; t on and t off is the minimum continuous duration of unit startup and shutdown; t0 is the current time t.

[0143] (3) Energy storage operation constraints, including electrochemical energy storage power station and pumped storage power station operation constraints, are expressed as follows:

[0144]

[0145] 0≤u b,dis (t)+u b,ch (t)≤1

[0146]

[0147] S b (1) = S b (N T )

[0148]

[0149] 0≤u vp (t)+u vg (t)≤1

[0150] V s min ≤V s (t)≤V s max

[0151] V s (1) = V s (N T )

[0152] In the formula, u b,ch (t),u b,dis (t) respectively represent the charging and discharging states of the electrochemical energy storage power station at time t; and They are the maximum charging power and maximum discharging power of the electrochemical energy storage power station respectively; and u is the upper and lower limits of the storage capacity of the electrochemical energy storage power station; vp (t),u vg (t) respectively represent the pumping and power generation status of the pumped storage power station at time t; P vp (t) and P vg (t) represent the pumping and generating power of the pumped storage power station at time t, respectively; They represent the upper and lower limits of the pumping power of the pumped storage power station respectively; Respectively represent the upper and lower limits of the pumping and power generation of the pumped storage power station; V s max With V s min The upper and lower limits of the water storage capacity of a pumped-storage power station.

[0153] (4) Power balance constraint, which is expressed as follows:

[0154]

[0155] (5) Line power flow and transmission capacity constraints, which are expressed as follows:

[0156]

[0157] Where: P iInject power into power system nodes; B ij 、x ij , P ij They are the branch susceptance, branch reactance and branch transmission power of the power system, among which, B ij =-1 / x ij , θ i ,θ j , Δθ ij They are respectively the phase angle of the node at the beginning of the branch, the phase angle of the node at the end of the branch, and the difference between the phase angles at the beginning and end of the branch.

[0158] Step 7: Use the Cplex solver to solve the two-stage optimization model and output the optimization scheduling result of the wind power system;

[0159] A. The Cplex optimization solver is used to solve the demand response upper optimization model and output optimized load information, including information such as the reducible load response amount, the transferable load response amount, the optimized load curve, and the demand response compensation cost.

[0160] B. Use the Cplex solver to solve the lower-level optimization model of the power system and output the optimization dispatch result information of the wind power system, including wind turbine output plan, traditional unit output plan, electrochemical energy storage power station output plan, pumped storage power station output plan, various energy consumption of the system, carbon emissions, economic operation costs and other information.

[0161] The present invention considers the low-carbon optimization scheduling method of the wind power system considering demand response. The calculation example takes the wind power system as the research object, the scheduling cycle is 24 hours, and the step length is 1 hour. Simulation verification is carried out based on the improved IEEE-30 node system: there are 6 thermal power units, located at nodes 1, 2, 5, 8, 11, and 13 respectively; there is 1 wind farm, located at node 2; the energy storage power station includes an electrochemical energy storage power station and a pumped storage power station, located at nodes 2 and 8 respectively; the electric loads are located at nodes 2, 3, 4, 7, 8, 10, 12, 14, 15, 16, 17, 18, 19, 20, 21, 23, 24, 26, 29, and 30 respectively. This chapter uses the MATLAB R2018b platform for preliminary data input, processing, and interface construction, and uses the general commercial optimization software CPLEX platform to write the model program and call the global solver for solution. Model program and call the global solver for solution.

[0162] The main parameter settings of the simulation example in this invention are as follows: Figure 2 and Figure 3The distribution probability information of each scenario is shown in Tables 1 and 2. The parameter information of the traditional generator set is shown in Table 3, and the parameter information of the energy storage device is shown in Table 4.

[0163] Table 1 Distribution probability of typical wind power output scenarios

[0164]

[0165] Table 2 Distribution probability of typical load demand scenarios

[0166]

[0167] Table 3 Equipment parameters of traditional thermal power units

[0168]

[0169]

[0170] Table 4 Energy storage equipment parameters

[0171]

[0172] In order to verify the effectiveness of the low-carbon optimization scheduling method of the wind power system considering demand response in the present invention, different scheduling schemes are set, Scheme 1: considering the conventional demand response incentive mechanism, the user response confidence is 100%; Scheme 2: considering the step-type demand response incentive mechanism, the user response confidence is 100%; Scheme 3: considering the improved step-type demand response incentive mechanism taking into account rewards and punishments, the user response confidence is 100%; Scheme 4: considering the improved step-type demand response incentive mechanism taking into account rewards and punishments, the user response confidence is 95%; Scheme 5: considering the improved step-type demand response incentive mechanism taking into account rewards and punishments, the user response confidence is 90%.

[0173] The main results and analysis of the simulation example of the low-carbon optimization scheduling method for wind power system considering demand response in the present invention: The system operation optimization results under three different scheduling schemes are shown in Table 5. By comparing the data in Table 5, it can be seen that Scheme 3 is obviously more advantageous than the first two schemes in terms of both economy and low carbon. The specific scheme comparison analysis is as follows.

[0174] In Scheme 1, users under the conventional demand response incentive mechanism are not very enthusiastic about responding, with a response rate of only 69.19%. The original peak-to-valley difference of the load curve was 0.6326, but after users participated in conventional demand response, the peak-to-valley difference of the load curve was only reduced by 0.0711. The optimization of the load structure on the demand side was not strong enough, resulting in a low matching degree between energy supply and demand. Although the system is equipped with an energy storage power station, this phenomenon still limits the system's use of low-cost, clean and pollution-free wind energy, and relies more on traditional thermal power units and energy storage power stations for regulation. Therefore, while the wind power consumption rate is low, carbon emissions are also high. Although the demand response cost is low, the comprehensive operating cost is relatively high, which leads to a high total cost of system optimization.

[0175] Scheme 2 adopts a stepped demand response incentive mechanism. Different from the single compensation price in Scheme 1, it guides users to participate in the response with multi-interval incentive prices, expanding the user's choice space. Compared with Scheme 1, the user response rate under Scheme 2 increased by 27.67%, the peak-to-valley difference rate of the load curve decreased by 0.0169, the total system operation cost decreased by US$5,490, the wind power consumption rate increased by 1.34%, and the carbon emissions decreased by 388 kg. Due to the significant peak-to-valley effect of the load curve under the stepped demand response incentive mechanism, the matching degree of the system supply and demand curves is greatly improved. The low-carbon economic performance of the system has been greatly improved mainly due to the optimization of the load structure: one reason is that although the compensation cost of demand response has increased by US$16,000 compared with Scheme 2, the operating cost of the system has been greatly reduced by US$21,490, and the cost reduction is far greater than the cost reduction, so the total optimization cost has been reduced. Another reason is that more wind power is absorbed and utilized by the system, replacing some traditional high-carbon thermal power units with the advantages of almost zero cost and no pollution, reducing the power generation cost in system operation, reducing carbon emissions and protecting the environment. These data strongly illustrate that the introduction of the step-type demand response incentive mechanism is of great significance for mobilizing users' enthusiasm for participating in the response, optimizing the load structure, and improving the low-carbon economic performance of system operation.

[0176] Scheme 3 adopts an improved step-type demand response incentive mechanism, and its user response rate is the same as that of Scheme 2, which is 96.86%, which once again verifies the outstanding role of the step-type demand response incentive mechanism in improving user participation enthusiasm compared with the conventional demand response incentive mechanism. Under the same high response rate, the peak-to-valley difference rate of the load curve of Scheme 3 is improved again, which is 0.0174 lower than that of Scheme 2. At the same time, the system wind power consumption rate increased by 0.23% and carbon emissions decreased by 15kg. These data show that in terms of improving the low-carbon nature of system operation and optimizing the load structure, Scheme 3 is not much different from Scheme 2 as a whole, and Scheme 3 is superior. In terms of the effect of improving the economic efficiency of system operation, Plan 3 demonstrates the superiority of the proposed improved stepped demand response incentive mechanism: under the condition of ensuring the same high response rate of users, Plan 3 further reduces the compensation cost of demand response by US$4,818, thereby reducing the total cost of system optimization by US$6,178, greatly improving the economic efficiency of system operation, and verifying the effectiveness and superiority of the improved stepped demand response incentive mechanism proposed in this chapter in optimizing load structure, promoting system wind consumption, and improving the economic and low-carbon operation of power systems.

[0177] Table 5 System operation optimization results under different scheduling schemes

[0178]

[0179]

[0180] Figure 4 This is a comparison chart of the load curves before and after the improved step-type demand response. The blue curve represents the original load demand curve, and the red curve represents the power curve after the user participates in the demand response to adjust his own power consumption behavior. It can be seen that the load curve after the demand response becomes flatter than the original load curve, and the difference between the maximum and minimum load values ​​is significantly reduced, which is reflected in the increase in load during the 1:00-5:00 and 23:00-24:00 periods, and the decrease in load during the 10:00-14:00 and 19:00-22:00 periods. This not only alleviates the pressure on energy supply during peak load periods, but also increases power demand during low load periods to help better absorb wind power with anti-peak characteristics.

[0181] Figure 5 It is the load optimization result under the improved step-type demand response incentive mechanism. The loads that users participate in responding include two types: curtailable loads and transferable loads. Figure 5The analysis shows that the improved step-type demand response can effectively achieve the effect of peak load reduction and valley filling. Part of the peak load is directly reduced by the reducible load, which reduces the total load demand during the peak period; the other part of the peak load is transferred from the peak period to the valley period, which realizes the flexible transfer of power load on the time axis and alleviates the contradiction between the supply and demand of insufficient peak supply and excess valley power in the system.

[0182] The optimal scheduling results of the system under Scheme 3 are as follows: Figure 6 As shown in the figure, the red curve is the optimized load curve after demand response. Wind power itself has the characteristics of anti-peak regulation, which is specifically manifested in that there is less wind power during peak load periods and more wind power during low load periods. The wind power output curve is exactly the opposite of the demand load curve under normal circumstances. In order to achieve the goal of optimal economic operation of the system, within the 24-hour dispatch cycle, the system will give priority to low-cost wind power for power generation under the two options of thermal power and wind power. In the night period of 1:00-3:00, there is more wind power and the load demand is not high. Wind power alone can fully achieve supply and demand balance. In the peak load period of 8:00-21:00, with the reduction of wind power output and the increase of load demand, wind power is almost completely absorbed by the system at this time, and the output of traditional thermal power units is gradually increasing, becoming the main force of system power supply during this period. During the remaining off-peak hours of 4:00-7:00 and 22:00-24:00, electricity is jointly supplied by traditional thermal power units and wind turbines. However, the wind energy is stronger at this time, so the output of thermal power units is reduced and the output of wind turbines accounts for a larger proportion.

[0183] Figure 7 The optimization scheduling results of the energy storage power station are shown. In the two periods of 1:00-5:00 and 23:00-24:00, the load is in a low demand state, and the excess wind power that cannot be used is stored in the energy storage equipment, so that the wind energy can be maximized, and the stored electricity is released in time during the peak period, relieving the energy supply pressure of the thermal power unit and reducing the cost of power generation. Since the cost of storing each megawatt of electricity in electrochemical energy storage is higher than that of pumped storage power stations, the system will give priority to dispatching pumped storage power stations to be in a charging state. Because the charging and discharging power of pumped storage power stations is limited by the maximum power value, at 10:00 and 22:00, the pumped storage power station reaches the maximum discharge power, and the electrochemical storage station is in a discharging state. Similarly, at 23:00 and 24:00, the pumped storage power station and the electrochemical energy storage power station are charged together to help absorb the excess wind power. It is precisely because of the intermediate regulating role of the energy storage power station that the system's wind power consumption is promoted, the flexible transfer of wind power on the time axis is realized, the efficiency of wind energy utilization is maximized, the operating cost of the system is effectively reduced, the pollution of carbon emissions to the environment is reduced, and the flexibility of system scheduling is improved.

[0184] The comparison of economic benefits of user response under three confidence levels is shown in Table 6. It can be seen that as the confidence level decreases, the benefits obtained by users participating in the response become smaller, which is specifically reflected in the reduction of demand response compensation income and the gradual increase in economic penalties imposed by the system on users for defaulting and not responding to loads. Under the improved step-type demand response incentive mechanism that takes into account rewards and penalties, users will inevitably try to avoid default penalties in order to maximize their own economic benefits, thereby increasing the user's actual response confidence and reducing the uncertainty risk of system operation. This verifies the important role of taking into account rewards and penalties in constraining user default behavior.

[0185] Table 6 Comparison of economic benefits of user response at different confidence levels

[0186]

[0187] The present invention takes the wind power system as the research object, proposes an improved step-type demand response incentive mechanism, constructs an upper-level optimization model for demand response and a lower-level optimization model for the power system, and proposes a low-carbon optimization scheduling method for the wind power system considering demand response, thereby improving the system's wind power absorption capacity and low-carbon economic performance of operation. The simulation results of the example of the present invention show that: (1) The improved step-type demand response incentive mechanism can effectively enhance the enthusiasm of users to participate in the response, has a better peak-shaving and valley-filling effect on the load, and improves the matching degree of the supply and demand curve. (2) Through the coordinated optimization of demand response and energy storage power stations, traditional units, and wind turbines, the absorption and utilization of wind power are promoted, which not only reduces the carbon emissions of the system, but also improves the economy of the system operation, achieving a win-win situation of economic and environmental benefits.

Claims

1. A low-carbon optimization scheduling method for a wind power system considering demand response, characterized in that: The steps include: Step 1: Collect the forecast data and forecast deviation information of the wind power output and load of the wind power system, and establish the uncertainty model of source and load, including the uncertainty model of wind power output and the uncertainty model of load demand; Step 2: Use Latin hypercube sampling method to randomly sample wind power and load respectively to obtain initial scenario sets of wind power and load; Step 3: Use the heuristic synchronous back-substitution method to reduce the initial scenario set of wind power and load, generate typical scenarios of wind power and load, and obtain the occurrence probability information of each typical scenario; Step 4: Establish a demand-side responsive load model, wherein the demand-side responsive load model includes a curtailable load model and a transferable load model; Step 5: Based on the established demand-side responsive load model, an improved step-type demand response incentive mechanism model is established; Step 6: Use a multi-scenario stochastic optimization method to build a two-stage optimization model, including a demand response upper-level optimization model and a power system lower-level optimization model; Step 7: Use the Cplex solver to solve the two-stage optimization model and output the optimization scheduling result of the wind power system.

2. The low-carbon optimization dispatching method for wind power system considering demand response according to claim 1, characterized in that: In step 1, the uncertainty model of source load is established as follows: (a) Uncertainty model of wind power output, which is expressed as follows: In the formula, They represent the actual output value and predicted output value of wind power at time t respectively; represents the wind power output prediction error; σ w (t) represents the standard deviation of wind power output prediction error; represents the total installed capacity of the wind farm; (b) The uncertainty model of load demand is expressed as follows: In the formula, and They represent the actual value and predicted value of the system load during period t respectively; represents the load forecast error; σ e (t) represents the standard deviation of load forecast error; e L It is expressed as a proportionality factor that characterizes the accuracy of the prediction tool.

3. The low-carbon optimization dispatching method for wind power system considering demand response according to claim 1, characterized in that: In step 2, the Latin hypercube sampling method is used to randomly sample wind power and load, respectively, and the initial scenario sets of wind power and load are obtained, which specifically includes the following steps: Step 21, set the power to follow the normal distribution N~(μ,δ), where μ is the expected value of the predicted value, δ is the fluctuation percentage, the number of samples is set to 1000, and the vertical axis of the cumulative probability distribution curve is divided into 1000 intervals with equal probability, and the width of each interval is 1 / 1000; Step 22: randomly select a value in each interval of the variable one by one, and select 1000 values ​​in total; Step 23: The cumulative distribution probability corresponding to the kth sampling value is F i (k)=(r u +k-1) / 1000, k=1,2,…,1000, where: r u ~N(0,1) obeys uniform distribution, and the kth sampling value of the random variable can be obtained through the inverse function; Step 24: After sampling, the sampling values ​​of all random variables are formed into a sampling matrix, and the sampling values ​​of each column of the matrix are sorted by using the Gram-Schemidt sequence orthogonalization method, and the correlation of each column is minimized through iterative calculation; Step 25: Form an initial scene set containing 1000 randomly sampled scenes.

4. The low-carbon optimization dispatching method for wind power system considering demand response according to claim 1, characterized in that: In step 3, the initial scenario set of wind power and load is reduced by using the heuristic synchronous back-substitution method to generate typical scenarios of wind power and load, and the occurrence probability information of each typical scenario is obtained, which includes the following steps: Step 31, define the number of target scenes to be reduced as 10, and the scene probability distance as the binary norm of the scene vector; Step 32: Calculate the Euclidean distance between any two scenes Step 33: Select the scene S i The closest scene S s , that is, D is =minD ij , calculate the probability distance p between scenes i and s is =D is p s , where D is It is scene S i and S s The Euclidean distance between s It is scene S s The probability of occurrence; Step 34: Find the probability distance value p in the sample is The smallest scene pair (S i ,S s ); Step 35: Let p s =p s +p i , in the update p s At the same time cut to scene S i , and correct the number of samples in the reduction process; Step 36, return to step 32 to perform iterative calculation, and stop the iteration when the number of scenes to be reduced is reached.

5. The low-carbon optimization dispatching method for wind power system considering demand response according to claim 1, characterized in that: In step 4, the demand-side responsive load model is established as follows: (a) The curtailable load model is expressed as follows: In the formula, ξ a (t), P a (t) They represent the load reduction state, load reduction amount, minimum load reduction amount, maximum load reduction amount and total load reduction amount at time t respectively; (b) Transferable load model, which is expressed as follows: In the formula, ξ bin (t) and ξ bout (t) are all 0-1 state variables, ξ bin (t) = 1 indicates load transfer, ξ bout (t) = 1 indicates load transfer; P bin (t), P bout (t) are the load transfer in and out at time t; P b (t) is the total load transfer at time t; are the upper and lower limits of the load transferred in at time t; is the upper and lower limits of the output load at time t; N T Indicates the scheduling period; (c) The demand side responsive load model is expressed as follows: In the formula, α represents the percentage of the total amount of defaulted and unresponsive load that cannot perform full response after the user signs the demand response contract to the total amount of signed response load; It indicates the total amount of flexible load actually provided by users in the demand response process during the dispatch period.

6. The low-carbon optimization dispatching method for wind power system considering demand response according to claim 1, characterized in that: In step 5, the improved step-type demand response incentive mechanism model is established as follows: (a) The reward mechanism model is expressed as follows: In the formula, f DR is the economic compensation for users participating in demand response; ε is the base price of compensation per unit of response electricity; v is the growth rate of the compensation unit price; d is the length of the incentive interval under the reward mechanism; (b) The penalty mechanism model is expressed as follows: In the formula, is the economic penalty for the user's failure to respond to the load, and its value is a negative number; τ represents the penalty base price per unit of default electricity; D represents the length of the step penalty interval; It represents the total load that is increased due to the user not fulfilling the response condition but going in the opposite direction.

7. The low-carbon optimization dispatching method for wind power system considering demand response according to claim 1, characterized in that: In step 6, the objective function of the upper-level optimization model of demand response is expressed as follows: In the formula, f1 represents the response benefit of users on the demand side; Represents the discomfort cost caused by users adjusting their electricity consumption behavior; ω a ,ω b are the unit electricity discomfort coefficients for load shedding and shifting, respectively; The objective function of the lower-level optimization model of the power system is expressed as follows: Where f2 represents the comprehensive operating cost of the system; ρ j represents the distribution probability of the jth typical scene; N j The total number of scenes representing typical scenes; f j,g represents the power generation cost of traditional thermal power units in the jth typical scenario; f j,ms represents the operating cost of the hybrid energy storage system; represents the carbon emission penalty cost generated by the traditional thermal power unit in the power generation process under the jth typical scenario; w b 、w s They are the operation and maintenance cost coefficients of the electrochemical energy storage power station and the pumped storage power station; P b (t), P s (t) is the output power of the electrochemical energy storage power station and the pumped storage power station at time t; is the carbon emission penalty cost coefficient; is the carbon emission coefficient per unit of electricity of the i-th thermal power unit; It represents the total carbon emissions of all thermal power units in the system at time t.

8. The low-carbon optimization dispatching method for wind power system considering demand response according to claim 7, characterized in that: The constraints of the lower-level optimization model of the power system include: A. Wind turbine operation constraints, the expressions are as follows: In the formula, is the upper limit of wind power output; B. Traditional unit operation constraints, including unit output upper and lower limit constraints, unit climbing constraints and minimum start and stop time constraints, are expressed as follows: In the formula, u g,i (t) and u g,i (t-1) represents the start and stop status of the i-th traditional unit at time t and time (t-1), respectively; They represent the minimum output power and maximum output power of the i-th conventional unit respectively; is the maximum climbing power of the traditional unit; t on and t off is the minimum continuous duration of unit startup and shutdown; t0 is the current time t; C. Energy storage operation constraints, including electrochemical energy storage power station and pumped storage power station operation constraints, are expressed as follows: 0≤u b,dis (t)+u b,ch (t)≤1 S b (1)=S b (N T ) 0≤u vp (t)+u vg (t)≤1 V s (1)=V s (N T ) In the formula, u b,ch (t),u b,dis (t) respectively represent the charging and discharging states of the electrochemical energy storage power station at time t; and They are the maximum charging power and maximum discharging power of the electrochemical energy storage power station respectively; and u is the upper and lower limits of the storage capacity of the electrochemical energy storage power station; vp (t),u vg (t) respectively represent the pumping and power generation status of the pumped storage power station at time t; P vp (t) and P vg (t) represent the pumping and generating power of the pumped storage power station at time t, respectively; They represent the upper and lower limits of the pumping power of the pumped storage power station respectively; They represent the upper and lower limits of the pumping and generating power of the pumped storage power station respectively; and The upper and lower limits of the water storage capacity of the pumped storage power station; D. Power balance constraint, its expression is as follows: E. Line flow and transmission capacity constraints, the expressions are as follows: Where: P i Inject power into power system nodes; B ij 、x ij , P ij They are the branch susceptance, branch reactance and branch transmission power of the power system, among which, B ij =-1 / x ij , θ i ,θ j , Δθ ij They are respectively the phase angle of the node at the beginning of the branch, the phase angle of the node at the end of the branch, and the difference between the phase angles at the beginning and end of the branch.

9. The low-carbon optimization dispatching method for wind power system considering demand response according to claim 1, characterized in that: In step 7, the Cplex optimization solver is used to solve the demand response upper optimization model, and the optimized load information is output, including the reducible load response amount, the transferable load response amount, the optimized load curve, and the demand response compensation cost information.

10. The low-carbon optimization dispatching method for wind power system considering demand response according to claim 1, characterized in that: The Cplex solver is used to solve the lower-level optimization model of the power system and output the optimization dispatch result information of the wind power system, including wind turbine output plan, traditional unit output plan, electrochemical energy storage power station output plan, pumped storage power station output plan, various energy consumption of the system, carbon emissions, and economic operation cost information.

Images