A time-of-use electricity price optimization method and terminal based on a scene set and electricity price elasticity

Abstract

The application discloses a time-of-use price optimization method based on a scenario set and price elasticity, constructs a price elasticity model based on the correlation between price and load; constructs a scenario set based on wind and light output and load through scenario generation and scenario reduction technology; under the scenario set, constructs a target function with the minimum wind and light abandonment rate and the maximum benefits of the power company and the load user as the target according to the price elasticity model, and constructs a time-of-use price optimization model; solves the time-of-use price optimization model, and performs time-of-use price optimization according to the solution result; the application uses the probability distribution of the scenario set to reflect the randomness of wind and light and load through scenario generation and scenario reduction technology, and proposes a time-of-use price optimization model with the minimum wind and light abandonment rate and the maximum benefits of the power company and the load user, effectively realizes the optimization of time-of-use price, and provides strong support for economic dispatching of the power system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electricity price optimization, in particular to a time-of-use electricity price optimization method and terminal based on a scenario set and electricity price elasticity. BACKGROUND

[0002] In recent years, with the increasing demand for energy worldwide, renewable energy has attracted much attention due to its cleanliness and sustainability. Renewable energy generation mainly includes wind power generation, photovoltaic power generation, pumped storage power generation, and geothermal power generation, which has been widely used in power systems. Due to the uncertainty of wind speed and light intensity, wind power output and photovoltaic power output have the disadvantages of uncertainty and randomness. With the increasing of wind power and photovoltaic power, the reliable operation of power systems is facing great challenges, and new requirements are put forward for economic dispatching. SUMMARY

[0003] The technical problem to be solved by the present application is to provide a time-of-use electricity price optimization method and terminal based on a scenario set and electricity price elasticity, which realizes the optimization of time-of-use electricity price.

[0004] To solve the above technical problems, the technical scheme adopted by the present application is:

[0005] A time-of-use electricity price optimization method based on a scenario set and electricity price elasticity, comprising the steps of:

[0006] S1, constructing an electricity price elasticity model based on the correlation between electricity price and load;

[0007] S2, constructing a scenario set based on wind and light output and load by scenario generation and scenario reduction technology;

[0008] S3, constructing a target function with the minimum wind and light abandonment rate and the maximum benefits of power companies and load users as the target under the scenario set, and constructing a time-of-use electricity price optimization model according to the electricity price elasticity model;

[0009] S4, solving the time-of-use electricity price optimization model and optimizing the time-of-use electricity price according to the solving result.

[0010] To solve the above technical problems, another technical scheme adopted by the present application is:

[0011] A time-of-use electricity price optimization terminal based on a scenario set and electricity price elasticity, comprising a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to realize the steps of the above-mentioned time-of-use electricity price optimization method based on a scenario set and electricity price elasticity.

[0012] The beneficial effects of the present application are that the time-of-use electricity price optimization method and terminal based on a scene set and electricity price elasticity, by scene generation and scene reduction technology, utilize the probability distribution of the scene set to reflect the randomness of wind, light and load, and propose a time-of-use electricity price optimization model that minimizes the comprehensive wind and light abandonment rate and maximizes the benefits of the power company and the load user, effectively realizing the optimization of time-of-use electricity price and providing strong support for economic dispatching of the power system. BRIEF DESCRIPTION OF DRAWINGS

[0013] Figure 1 A flowchart of a time-of-use electricity price optimization method based on a scene set and electricity price elasticity according to an embodiment of the present application;

[0014] Figure 2 A structural diagram of a time-of-use electricity price optimization terminal based on a scene set and electricity price elasticity according to an embodiment of the present application;

[0015] Figure 3 A diagram showing the relationship between electricity price and load of a time-of-use electricity price optimization method based on a scene set and electricity price elasticity according to an embodiment of the present application;

[0016] Figure 4 A flowchart of a K-means clustering algorithm of a time-of-use electricity price optimization method based on a scene set and electricity price elasticity according to an embodiment of the present application;

[0017] Figure 5 A flowchart of an MOPSO optimization algorithm of a time-of-use electricity price optimization method based on a scene set and electricity price elasticity according to an embodiment of the present application;

[0018] Figure 6 A diagram showing the wind power output of 365 scenes of a time-of-use electricity price optimization method based on a scene set and electricity price elasticity according to an embodiment of the present application;

[0019] Figure 7 A diagram showing the wind power output under a scene set of a time-of-use electricity price optimization method based on a scene set and electricity price elasticity according to an embodiment of the present application;

[0020] Figure 8 A diagram showing the photovoltaic output power of 365 scenes of a time-of-use electricity price optimization method based on a scene set and electricity price elasticity according to an embodiment of the present application;

[0021] Figure 9 A diagram showing the photovoltaic output power under a scene set of a time-of-use electricity price optimization method based on a scene set and electricity price elasticity according to an embodiment of the present application;

[0022] Figure 10 A diagram showing the load of 365 scenes of a time-of-use electricity price optimization method based on a scene set and electricity price elasticity according to an embodiment of the present application;

[0023] Figure 11 A load diagram under a scenario set of a time-of-use price optimization method based on a scenario set and price elasticity according to an embodiment of the present application;

[0024] Figure 12 A load distribution diagram of a peak-valley period in five cases of a time-of-use price optimization method based on a scenario set and price elasticity according to an embodiment of the present application;

[0025] Figure 13 An MOPSO simulation result diagram of a multi-objective time-of-use price optimization model of a time-of-use price optimization method based on a scenario set and price elasticity according to an embodiment of the present application;

[0026] Label explanation:

[0027] 1. A time-of-use price optimization terminal based on a scenario set and price elasticity; 2. A processor; 3. A memory. DETAILED DESCRIPTION

[0028] To make the technical content, the purposes and effects of the present application clear, the following will be described in detail in combination with the embodiments and the accompanying drawings.

[0029] Please refer to Figure 1 A time-of-use price optimization method based on a scenario set and price elasticity, comprising the steps of:

[0030] S1. Constructing a price elasticity model based on the correlation between price and load;

[0031] S2. Constructing a scenario set based on wind and light output and load through scenario generation and scenario reduction technology;

[0032] S3. Under the scenario set, constructing a target function with the minimum wind and light abandonment rate and the maximum benefits of the power company and the load user as the target according to the price elasticity model, and constructing a time-of-use price optimization model;

[0033] S4. Solving the time-of-use price optimization model, and performing time-of-use price optimization according to the solution result.

[0034] From the above description, the beneficial effects of the present application are that: a time-of-use price optimization method based on a scenario set and price elasticity, through scenario generation and scenario reduction technology, uses the probability distribution of the scenario set to reflect the randomness of wind and light and load, and proposes a time-of-use price optimization model that minimizes the comprehensive wind and light abandonment rate and maximizes the benefits of the power company and the load user, effectively realizing the optimization of time-of-use price, and providing strong support for economic dispatching of the power system.

[0035] Further, step S1 is specifically:

[0036] According to the relationship curve between electricity price and user's load, the electricity price elasticity of the load is initially defined as:

[0037]

[0038] wherein L0(t) is the initial load at time t, P0(t) is the initial electricity price at time t, ΔP(t) is the electricity price change amount at time t, and ΔL(t) is the load change amount at time t;

[0039] The electricity price elasticity model using a linear function is expressed as:

[0040] L(t) = a + b·P(t);

[0041] wherein a and b are linear function coefficients of the load to the electricity price;

[0042] The electricity price elasticity of the load can be calculated as:

[0043]

[0044] According to the influence of the valley electricity price of the time-of-use electricity price on the load in the flat time period and the peak time period, the electricity price elasticity matrix is expressed as:

[0045]

[0046] wherein E P-O , E P-V , and E O-V are cross elasticity coefficients of the peak time period and the flat time period, the peak time period and the valley time period, and the flat time period and the valley time period, respectively, E P-P , E V-V , and E O-O are self elasticity coefficients of the peak time period, the valley time period, and the flat time period, respectively, and are calculated from the electricity price elasticity calculation formula of the load;

[0047] The load L1 after the electricity price is modified is expressed as:

[0048]

[0049] wherein L P,0 , L O,0 , and L V,0 are the initial loads of the peak time period, the flat time period, and the valley time period, respectively, L0 = [L P, 0L O,0 L V,0 ] is the load under the initial electricity price, and ΔL P , ΔL O , and ΔL V are the load change amounts of the peak time period, the flat time period, and the valley time period, respectively.

[0050] From the above description, by the above steps, the construction of the electricity price elasticity model is realized, compared with the traditional electricity price elasticity matrix, the electricity price elasticity model of the linear function is introduced, which can better reflect the close relationship between the load and the electricity price.

[0051] Further, step S2 comprises the steps of:

[0052] S21, selecting wind and light output and load data of a typical day for construction;

[0053] S22, based on the wind and light output and load data of the typical day, a first preset number of load and wind and light output scenarios are randomly generated by using the hyper-Latin cube sampling technology;

[0054] S23, by scene reduction technology, a second preset number of load and wind and light output scenarios are determined from the first preset number of load and wind and light output scenarios, and a scene set is obtained.

[0055] From the above description, in the optimization of time-of-use price research, if only the typical day mode of wind and light output and load data is considered, the influence of the volatility of load and wind and light output on the optimization result will be ignored, therefore, the present application takes the wind and light output and load data of the typical day as the basis, constructs the scene set of wind and light output and load, and uses the probability distribution of the scene set to reflect the randomness of wind and light and load.

[0056] Further, step S21 is specifically selected by K-means clustering algorithm to construct wind and light output and load data of a typical day.

[0057] From the above description, the wind and light output and load data of a typical day are selected by the K-means clustering algorithm for construction as a specific embodiment of the present application.

[0058] Further, step S21 comprises the steps of:

[0059] S211, obtaining M days of load and wind and light output data generated by sampling;

[0060] S212, initializing the load and wind and light output data, and randomly extracting k days of load and wind and light output data as the clustering centers of various data, setting the maximum number of loops and the change error of the clustering centers;

[0061] S213, calculating the Euclidean distance between each day's sample data in the load data and the wind and light output data and the k clustering centers;

[0062] S214, taking each clustering center as a class, and classifying each sample data according to the minimum Euclidean distance;

[0063] S215, recalculating the clustering center of data in each category;

[0064] S216, judging whether the clustering center changes or whether the maximum number of times is reached, if the clustering center does not change or the maximum number of iterations is reached, the clustering result is obtained and step S217 is entered, otherwise step S213 is returned;

[0065] S217, according to the comparison of the clustering center distance of the clustering result, the typical scene corresponding to the data is obtained.

[0066] From the above description, it can be known that based on the above steps, the selection of typical day wind power output and load data is realized by K-means clustering algorithm.

[0067] Further, the initialization of the load and wind power output data in step S212 is specifically:

[0068] The load and wind power output data to be classified are standardized by z-score standardization processing:

[0069]

[0070] Wherein, Z m,t is the standardized data value of the mth day at t time, μ t is the mean value of the overall data at t time, δ t is the standard deviation of the overall data at t time.

[0071] From the above description, it can be known that since K-means clustering is based on distance as the classification standard, standardizing the data to be classified can avoid inaccurate classification caused by too large variable difference.

[0072] Further, the objective function of the time-of-use electricity price optimization model, which takes the maximum benefit f1 of the power company and the load user as the target, is represented as:

[0073]

[0074] Wherein, Rcom l is the benefit of the power company under l scene, Ruser l is the benefit of the load user under l scene, α and β are the weight coefficients of the benefits of the power company and the load user respectively, p l is the probability of l scene, and the relationship between α and β is α+β=1;

[0075] The benefit Rcom of the power company includes the income of selling electricity to users and the cost of purchasing electricity from the large power grid, which is represented as:

[0076]

[0077] wherein P(t) is the electricity price at time t, L1(t) is the load at time t, a buy P(t) is the electricity price at time t, L1(t) is the load at time t, a buy P(t) is the electricity price at time t, L1(t) is the load at time t, a

[0078] The income Ruser of the load user is expressed as:

[0079]

[0080] wherein C0 is the electricity cost of the load user under the initial electricity price, C TOU is the electricity cost of the load user calculated according to the time-of-use electricity price, P0(t) is the initial electricity price at time t, L0(t) is the initial load at time t, and L1(t) is the load after the electricity price is corrected.

[0081] From the above description, it can be known that based on the above steps, the construction of the objective function with the goal of maximizing the income of the power company and the load user is realized.

[0082] Further, the objective function with the goal of minimizing the curtailment rate f2 in the time-of-use electricity price optimization model is expressed as:

[0083]

[0084] wherein P char (t) is the charging power of the ESS at time t, P PV.upper (t) and P wind.upper (t) are the maximum value of the photovoltaic power and the maximum value of the wind power at time t respectively.

[0085] From the above description, it can be known that based on the above steps, the construction of the objective function with the goal of minimizing the curtailment rate is realized.

[0086] Further, the constraint conditions in the time-of-use electricity price optimization model include:

[0087] Power balance constraint:

[0088] The power balance requirement needs to be met for the wind and light power generation, the charging and discharging power of the energy storage system, and the load, and is expressed as:

[0089] L1(t) + P char (t) = P PV (t) + P wind (t) + P dis (t) + d buy (t);

[0090] wherein L1(t) is the load after the electricity price is corrected, P char(t) is the charging power of the ESS at time t, P PV (t) is the photovoltaic output power at time t, P wind (t) is the wind power output at time t, P dis (t) is the discharging power of the ESS at time t;

[0091] Photovoltaic and wind power constraints:

[0092] Wind and photovoltaic power are mainly determined by the maximum and minimum values of wind and photovoltaic power, expressed as:

[0093] 0≤P PV (t)≤P PV.upper (t);

[0094] 0≤P wind (t)≤P wind.upper (t);

[0095] where P PV.upper (t) is the maximum photovoltaic power at time t, P wind.upper (t) is the maximum wind power at time t;

[0096] ESS constraints:

[0097] When the ESS is charging, the battery pack is powered by the surplus wind and photovoltaic output, and the storage energy of the battery is determined by the charging power and the storage energy of the previous hour, expressed as:

[0098] E(t) = E(t-1) + P char (t) · η char ;

[0099] where E(t) is the storage energy of the battery at time t, and η char is the charging efficiency of the battery;

[0100] When the ESS is discharging, the battery pack powers the load, and the storage energy of the battery is determined by the discharging power and the storage energy of the previous hour, expressed as:

[0101] E(t) = E(t-1) - P dis (t) / η dis ;

[0102] where P dis is the discharging power of the battery, and η dis is the discharging efficiency of the battery;

[0103] The SOC constraint is expressed as:

[0104] S OC.lower ≤ S OC (t) ≤ SOC.upper ;

[0105] wherein S OC (t) is the SOC of the ESS at time t, S OC.lower denotes the minimum state of charge of the ESS, S OC.upper denotes the maximum state of charge of the ESS;

[0106] The discharge power constraint of the ESS is expressed as:

[0107] 0 < P dis (t) < min(P dis.upper ,(S OC (t-1)-S OC.lower )E ESS );

[0108] The charge power constraint of the ESS is expressed as:

[0109] 0 < P char (t) < min(P char.upper ,(S OC.upper -S OC (t-1))E ESS );

[0110] wherein P char.upper and P dis.upper are the maximum values of the charge power and the discharge power of the ESS, respectively, and E ESS is the rated capacity of the battery pack.

[0111] As described above, the time-of-use price optimization model further includes a plurality of constraint conditions such as a power balance constraint, a photovoltaic and wind power constraint, and an ESS constraint.

[0112] Please refer to Figure 2 , a time-of-use price optimization terminal based on a scenario set and price elasticity, comprising a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the time-of-use price optimization method based on the scenario set and the price elasticity described above when executing the computer program.

[0113] As described above, the time-of-use price optimization terminal based on the scenario set and the price elasticity has the beneficial effects that: through scenario generation and scenario reduction technology, the probability distribution of the scenario set is used to reflect the randomness of wind, light and load, and a time-of-use price optimization model that minimizes the comprehensive wind and light rejection rate and maximizes the benefits of the power company and the load user is proposed, thereby effectively realizing the optimization of the time-of-use price and providing strong support for the economic dispatching of the power system.

[0114] The application discloses a time-of-use electricity price optimization method and terminal based on a scenario set and electricity price elasticity, and is suitable for time-of-use electricity price optimization of a power grid system, and in particular for time-of-use electricity price optimization of a micro-grid system composed of wind power generators, photovoltaic power generators and energy storage systems (ESS).

[0115] Please refer to Figure 1 and Figures 3 to 5 Embodiment one of the application is:

[0116] A time-of-use electricity price optimization method based on a scenario set and electricity price elasticity comprises the following steps:

[0117] S1, constructing an electricity price elasticity model based on the correlation between electricity prices and loads;

[0118] Step S1 is specifically:

[0119] According to the relationship curve between electricity prices and user loads, the electricity price elasticity of the load is initially defined as:

[0120]

[0121] Wherein, L0(t) is the initial load at time t, P0(t) is the initial electricity price at time t, ΔP(t) is the electricity price change at time t, and ΔL(t) is the load change at time t.

[0122] The electricity price elasticity model is expressed by using a linear function as:

[0123] L(t) = a + b·P(t) (2)

[0124] Wherein, a and b are linear function coefficients of the load to the electricity price.

[0125] Then, the electricity price elasticity of the load can be calculated as:

[0126]

[0127] According to the influence of the valley electricity price of the time-of-use electricity price on the loads in the flat time period and the peak time period, the electricity price elasticity matrix is expressed as:

[0128]

[0129] Wherein, E P-O , E P-V , E O-V are respectively the cross elasticity coefficient of the peak time period and the flat time period, the cross elasticity coefficient of the peak time period and the valley time period, and the cross elasticity coefficient of the flat time period and the valley time period, E P-P , E V-V , E O-OSelf-elasticity coefficients of peak time period, valley time period and flat time period, calculated by the formula of load price elasticity;

[0130] For example, the expression of cross-elasticity coefficient of peak time period and flat time period is:

[0131]

[0132] Wherein, b O , b P , a P are linear coefficient of flat time period and linear coefficient of peak time period respectively, P O , P P are price of flat time period and peak time period respectively. The other two cross-elasticity coefficients are similar to this formula.

[0133] The load L1 after the price is modified is represented as:

[0134]

[0135] Wherein, L P,0 , L O,0 , L V,0 are initial load of peak time period, flat time period and valley time period respectively, L0=[L P, 0L O,0 L V,0 ] is load under initial price, ΔL P , ΔL O , ΔL V are load change of peak time period, flat time period and valley time period respectively.

[0136] In this fact example, from the economic point of view, the load is closely related to the price. Therefore, on the basis of meeting the basic load demand, the user adjusts the load according to the time-of-use price, and minimizes the cost of using electricity. In addition to the basic load of user production and life, the user can adjust the daily load curve according to the price. For example, the user charges the electric vehicle, increases the charging during the period when the price is low, and reduces the charging during the period when the price is high. The relationship between the price and the load is shown in Figure 3 It can be seen from Figure 3 that the load of the user decreases with the increase of the price. In this fact example, four kinds of functional relationships between the load and the price are proposed, including linear function, potential function, logarithmic function and exponential function. In this embodiment, the time-of-use price optimization model is constructed by using the price elasticity of linear function.

[0137] The construction of the price elasticity model can be realized by referring to the above steps.

[0138] S2, a scene set based on wind and light output and load is constructed by scene generation and scene reduction technology;

[0139] Step S2 comprises steps of:

[0140] S21, selecting wind and light output and load data of a typical day to construct;

[0141] Step S21 is specifically selecting wind and light output and load data of a typical day by K-means clustering algorithm, comprising steps of:

[0142] S211, obtaining M days of load and wind and light output data generated by sampling;

[0143] S212, initializing the load and wind and light output data, and randomly extracting k days of load and wind and light output data as clustering centers of various data, setting a maximum number of iterations and a change error of the clustering centers;

[0144] The initialization of the load and wind and light output data in step S212 is specifically:

[0145] The load and wind and light output data to be classified are standardized by z-score standardization processing:

[0146]

[0147] Wherein, Z m,t is the standardized data value of the mth day at t time, μ t is the mean value of the overall data at t time, δ t is the standard deviation of the overall data at t time;

[0148] S213, calculating the Euclidean distance between each day's sample data and the k clustering centers in the load data and wind and light output data.

[0149] Euclidean distance is one of the clustering bases of K-means algorithm, and the formula is used to calculate the distance from each sample data to each clustering center, and its expression is shown in formula (7):

[0150]

[0151] In the formula, is the standardized sample data value of the mth day; K = (K n1 ,K n2 ,…,K n24 ) is the nth clustering center, n is a positive integer, when clustering for the first time, n clustering centers are randomly specified in the sample data, and in the process of each iteration, the clustering center K n is also updated.

[0152] S214, each clustering center is taken as a class, and each sample data is classified according to the minimum Euclidean distance.

[0153] In this embodiment, the sample data of the mth day is divided into K ω , whose expression is shown as formula (8):

[0154] d ata,m ∈K w ,D(d ata,m ,K w )=min{D(d ata,m ,K1),…D(d ata,m ,K w ),…,D(d ata,m ,K n )};(8)

[0155] In the formula, ω is a positive integer satisfying 1≤ω≤n.

[0156] S215, recalculating the clustering center of data in each class.

[0157] In this embodiment, the new clustering center of each class is calculated again according to the classification. When the clustering center no longer changes, it indicates that the optimal clustering state has been reached. The expression of the clustering center is shown as formula (9) and (10):

[0158]

[0159] K' n =(K' n1 ,K' n2 ,…,K' n24 ); (10)

[0160] In the formula, j x is the number of samples in the nth class; z p1 and z p24 are the pth data at 1 and 24 o'clock respectively.

[0161] The reduced scenarios should satisfy:

[0162]

[0163] In the formula, p(l) represents the probability of the lth scenario, and L represents the number of reduced scenarios.

[0164] S216, judging whether the clustering center changes or whether the maximum number of iterations is reached. If the clustering center does not change or the maximum number of iterations is reached, the clustering result is obtained and step S217 is entered, otherwise step S213 is returned;

[0165] S217, obtaining the typical scenario of the corresponding data according to the comparison of the clustering center distance of the clustering result.

[0166] In this fact example, the wind and light output and load data of the typical day are selected by the K-means clustering algorithm. The K-means clustering algorithm is widely used due to its simple principle and fast convergence speed. In order to make the selected typical day more representative, the K-means clustering algorithm is used to cluster the wind speed, light radiation intensity, etc. in this embodiment. The K-means clustering algorithm process is as shown in Figure 4 .

[0167] S22, based on the wind and light output and load data of the typical day, a first preset number of load and wind and light output scenarios are randomly generated by using the super-Latin cube sampling technology;

[0168] S23, a second preset number of load and wind and light output scenarios are determined from the first preset number of load and wind and light output scenarios by using the scene reduction technology, and a scene set is obtained.

[0169] In this embodiment, it is considered that in the study of optimal time-of-use electricity price, if only the typical day mode of wind and light output and load data is considered, the influence of the volatility of load and wind and light output on the optimization result will be ignored. Therefore, in this embodiment, the wind and light output and load data of the typical day are used as the basis to construct the scene set of wind and light output and load, and it is assumed that the error of these load and wind and light output data conforms to the normal distribution. By using the super-Latin cube sampling technology, M load and wind and light output scenarios are randomly generated, and by using the scene reduction technology, a scene set of L load and wind and light output scenarios is finally determined.

[0170] S3, under the scene set, a target function with the minimum wind and light rejection rate and the maximum benefits of the power company and the load user is constructed according to the electricity price elasticity model, and a time-of-use electricity price optimization model is constructed;

[0171] In this embodiment, the utilization rate of wind power and photovoltaic output and the economy of the microgrid are taken as two important indicators. In the established time-of-use electricity price optimization model, the target function includes the benefits of the power company and the load user and the wind and light rejection rate. It should be noted that this embodiment mainly analyzes the relationship between the time-of-use electricity price and the cost of the power company and the cost of the load user. However, there is no correlation between the time-of-use electricity price and the wind power generation and the photovoltaic power generation. The cost of wind power generation and photovoltaic power generation is not considered in this embodiment, and the cost of the energy storage system is also not considered. The microgrid in this embodiment is only composed of photovoltaic power generation, wind power generation and ESS, and does not include traditional energy generation methods.

[0172] In this embodiment, the time-of-use electricity price optimization model aims to find the optimal time-of-use electricity price. The multi-objective function established includes: the benefits of the power company and the load user; the wind and light rejection rate.

[0173] The time-of-use pricing optimization model proposed in this embodiment considers the revenue of both the power company and the load users. However, depending on the specific circumstances, optimizing the time-of-use pricing has a direct impact on the revenue of both the power company and the load users. The power company's revenue includes electricity sales revenue and expenses incurred in purchasing electricity from the main grid. The load user's revenue is the cost savings in electricity charges after implementing the optimal time-of-use pricing, which is the electricity cost at the initial price minus the electricity cost after implementing the optimal time-of-use pricing.

[0174] The objective function in the time-of-use pricing optimization model, which aims to maximize the revenue f1 for both the power company and the load user, is expressed as follows:

[0175]

[0176] Among them, Rcom l For the power company's revenue in scenario l, Ruser l For the revenue of load users in scenario l, α and β are the weighting coefficients of the revenue of the power company and the load user, respectively, and p l Let α be the probability of scenario l occurring, and the relationship between α and β is α + β = 1;

[0177] The revenue of a power company, Rcom, includes the revenue from selling electricity to customers and the cost of purchasing electricity from the main grid, expressed as:

[0178]

[0179] Where P(t) is the electricity price at time t, L1(t) is the load at time t, and a buy Let (t) be the electricity price purchased from the main power grid at time t, and d buy (t) represents the amount of electricity purchased from the main power grid at time t.

[0180] For users, electricity cost is the primary consideration. However, users' electricity costs are mainly affected by electricity prices and load.

[0181] The revenue (Ruser) of a load user is represented as:

[0182]

[0183] Where C0 is the electricity cost for the load user under the initial electricity price, C TOU This is the electricity cost of load users calculated according to time-of-use pricing. P0(t) is the initial electricity price at time t, L0(t) is the initial load at time t, and L1(t) is the load after adjusting the electricity price.

[0184] Because the wind power and photovoltaic power have the characteristics of non-scheduling, when the wind power and photovoltaic power exceed the load, the excess output power of the wind power and photovoltaic power can only be discarded. By optimizing the time-of-use electricity price, the load curve can be close to the sum of the wind power and photovoltaic power, so as to reduce the rate of abandoned wind power and photovoltaic power.

[0185] The objective function with the minimum abandoned wind power and photovoltaic power rate f2 as the target in the time-of-use electricity price optimization model is represented as:

[0186]

[0187] Wherein, P char (t) is the charging power of the ESS at time t, P PV.upper (t) is the maximum value of the photovoltaic power at time t, and P wind.upper (t) is the maximum value of the wind power at time t.

[0188] The constraint conditions in the time-of-use electricity price optimization model include:

[0189] Power balance constraint:

[0190] The power balance requirements of the wind power and photovoltaic power, the charging and discharging power of the energy storage system, and the load need to be met, and are represented as:

[0191] L1(t) + P char (t) = P PV (t) + P wind (t) + P dis (t) + d buy (t); (16)

[0192] Wherein, L1(t) is the load after the electricity price is corrected, P char (t) is the charging power of the ESS at time t, P PV (t) is the photovoltaic output power at time t, P wind (t) is the wind power output power at time t, and P dis (t) is the discharging power of the ESS at time t.

[0193] Photovoltaic and wind power constraint:

[0194] The wind power and photovoltaic power generation in the microgrid is mainly determined by the maximum and minimum values of the wind power and photovoltaic output power. The utilization of the photovoltaic power and wind power should meet the following constraint conditions:

[0195] 0 ≤ P PV (t) ≤ P PV.upper (t); (17)

[0196] 0 ≤ P wind (t) ≤ P wind.upper(t); (18)

[0197] Among them, P PV.upper (t) represents the maximum photovoltaic power at time t, P wind.upper (t) represents the maximum wind power at time t;

[0198] ESS constraints:

[0199] During the charging process of the ESS, the battery pack is powered by surplus wind and solar power. The stored energy of the battery is determined by the charging power and the stored energy in the previous hour, as follows:

[0200] E(t)=E(t-1)+P char (t)·η char (19)

[0201] Where E(t) is the energy stored in the battery at time t, and η char The charging efficiency of the battery;

[0202] When the ESS is discharging, the battery pack supplies power to the load. The stored energy of the battery is determined by the discharge power and the stored energy of the previous hour, expressed as:

[0203] E(t)=E(t-1)-P dis (t) / η dis (20)

[0204] Among them, P dis η is the discharge power of the battery. dis The discharge efficiency of the battery;

[0205] State of charge (SOC) reflects the remaining charge of a battery. The SOC value is a relative value, expressed as a percentage. The SOC value is 0 ≤ SOC ≤ 100%, where SOC = 100% indicates the battery is fully charged, and SOC = 0% indicates the battery is fully discharged. In this embodiment, the self-discharge of the energy storage device (ESS) is ignored, and the SOC constraint is expressed as follows:

[0206] S OC.lower ≤S OC (t)≤S OC.upper ; (twenty one)

[0207] Among them, S OC (t) is the SOC of ESS at time t, S OC.lower S represents the minimum state of charge of the ESS. OC.upper This indicates the maximum state of charge of the ESS;

[0208] The discharge power constraint of the ESS is expressed as:

[0209] 0≤P dis (t)≤min(P dis.upper ,(S OC (t-1)-S OC.lower E ESS ); (twenty two)

[0210] The charging power constraint of the ESS is expressed as follows:

[0211] 0≤P char (t)≤min(P char.upper ,(S OC.upper -S OC (t-1))E ESS ); (twenty three)

[0212] Among them, P char.upper and P dis.upper These are the maximum charging power and the maximum discharging power of the ESS, respectively. ESS This refers to the rated capacity of the battery pack.

[0213] In the time-of-use pricing optimization model proposed in this embodiment, the time-of-use price is used as the optimization variable. Based on hourly load, the day is divided into peak, valley, and flat periods. In this embodiment, the initial electricity price is set at 535.72 yuan / MWh. The three time periods for the time-of-use price are shown in Table 1. In the electricity market of the power system, in order to maintain price stability, the electricity price of the power company should be adjusted within a certain range. The constraints of the time-of-use price are shown in Table 2.

[0214] Table 1. Time-of-use electricity pricing

[0215]

[0216] Table 2. Time-of-use pricing constraints

[0217]

[0218] S4. Solve the time-of-use electricity pricing optimization model and optimize the time-of-use electricity pricing based on the solution results.

[0219] Please refer to Figure 5 Embodiment two of the present invention is as follows:

[0220] A time-of-use electricity pricing optimization method based on scenario sets and electricity price elasticity is presented in this embodiment, which explains how to solve the time-of-use electricity pricing optimization model.

[0221] Genetic Algorithm (GA) and Multiple Objective Particle Swarm Optimization (MOPSO) are mature optimization algorithms that have been applied by many scholars.

[0222] In this embodiment, MOPSO is used to solve the multi-objective time-of-use pricing optimization model. A satisfactory solution is selected from the obtained Pareto optimal solution set.

[0223] The standardized objective function for each solution is expressed as:

[0224]

[0225] Where s(x) = [f1(x), f2(x)] is the objective function, and f = [f1, f2] is the standardized objective function; f 1.max and f 2.max Let f be the maximum value of objective function f1 and function f2, respectively; 1.min and f 2.min These are the minimum values ​​of objective functions f1 and f2, respectively.

[0226] The normalized function value of the non-dominated solution can be calculated as follows:

[0227]

[0228] In the formula, N σ，i Let N be the normalized function value with i non-dominated solutions; N is the number of objective functions; and M is the number of non-dominated solutions. σ，i The maximum value determines the optimal solution of the multi-objective function. Figure 5 The flowchart shown is the MOPSO solution flowchart for the time-of-use electricity pricing optimization model.

[0229] (1) Input initial data such as load, electricity price, wind power, and photovoltaic.

[0230] (2) Randomly generate variables (time-of-use electricity price) within the time-of-use electricity price limit.

[0231] (3) The load under random time-of-use pricing is calculated according to the formula in the electricity price elasticity model.

[0232] (4) Compare the load with the wind power and photovoltaic output to determine the charging or discharging process of the ESS. If the wind power and photovoltaic output can meet the load, the ESS will operate during the charging process. If the wind power and photovoltaic output cannot meet the load, the ESS will operate during the discharging process. The multi-objective functions f1 and f2 are calculated by equations (12)-(15).

[0233] (5). The non-dominated solution set is obtained from step (4).

[0234] (6) Update time-of-use electricity pricing and speed.

[0235] (7). Update the new solution using the solution from mutation step (6) and determine the new non-dominated solution.

[0236] (8). Stopping Criteria. When the number of iterations is less than the number parameter set in this study, proceed to step (6); otherwise, stop according to the stopping criterion.

[0237] (9) The Pareto optimal set is obtained by applying the multi-objective particle swarm optimization algorithm (MOPSO).

[0238] (10). From the formula:

[0239]

[0240] Japanese style:

[0241]

[0242] A satisfactory solution was obtained.

[0243] Embodiment 3 of the present invention is as follows:

[0244] A time-of-use electricity pricing optimization method based on scenario sets and electricity price elasticity is presented in this embodiment. The microgrid mainly consists of a photovoltaic power station, a wind power station, and an energy storage system. The parameters of the storage battery are shown in Table 3.

[0245] Table 3. Battery Parameters

[0246]

[0247]

[0248] In the linear function coefficients of the electricity price elasticity model, the values ​​of a and b for peak, flat, and valley periods are 300 and -1.5, 245 and -3.5, and 198 and -4.5, respectively. These data are taken from the literature "YOUSEFI S, MOGHADDAM MP, MAJD VJ. Optimal real-time pricing in an agent-based retail market using a comprehensive demand response model[J]. Energy, 2011, 36(9): 5716-27." with some modifications. Since the large power grid also experiences supply pressure during peak periods, the electricity price purchased by the microgrid from the large power grid varies across different time periods. In the example of this study, the electricity price purchased from the large power grid during peak, flat, and valley periods is 928.2 yuan / MWh, 571.2 yuan / MWh, and 285.5 yuan / MWh, respectively.

[0249] In this embodiment, the MOPSO algorithm is used to solve the proposed time-of-use electricity pricing optimization model. In the MOPSO optimization algorithm, the population size is 200, and the number of iterations is 1000. The optimization algorithm is run on a computer with an i5-1240P CPU and 16GB of RAM using MATLAB 2019 software.

[0250] The basic data for wind power, photovoltaic power, and load are derived from the literature "NOJAVAN S, ZARE K, MOHAMMADI-IVATLOOB. Optimal stochastic energy management of retailer based on selling price determination under smart grid environment in the presence of demand response program[J]. Applied Energy, 2017, 187: 449-64.", with some modifications. Using scene generation and reduction techniques, the discrete probability distributions of wind power, photovoltaic output power, and load are shown in Table 4.

[0251] Table 4. Probability distribution of the reduced scene set

[0252]

[0253] Among them, the wind power, photovoltaic output power and load of the 365 generated scenarios are as follows: Figure 6 , Figure 8 and Figure 10As shown. The reduced wind and solar power output and load of the constructed scenario set are as follows: Figure 7 , Figure 9 and Figure 11 As shown.

[0254] In this embodiment, MOPSO is used to solve five cases to verify the effectiveness of the time-of-use pricing optimization model proposed in this study. Different weighting coefficients are set for the revenue of power companies and load users in the five cases. The optimization results are shown in Tables 5, 6, and 7, yielding optimized time-of-use pricing, revenue for power companies and load users, and wind and solar curtailment rates. Figure 12 The bar chart shows the load distribution during peak, off-peak, and valley periods under the benchmark electricity price and the optimized electricity prices for five case studies.

[0255] Table 5. Optimization results and objective function values ​​for the five cases.

[0256]

[0257] Table 6. Detailed list of optimization results for the five cases.

[0258]

[0259] Table 7. Load Distribution in Different Periods

[0260]

[0261] The simulation results of solving the time-of-use pricing optimization model are as follows: Figure 13 As shown. By Figure 13 It can be seen that there is a certain contradiction between the overall revenue of power companies and load users and the wind and solar curtailment rate. The optimized electricity price during off-peak and flat periods is lower than the benchmark case, while the optimized electricity price during peak periods is higher. This reduces the amount of excess electricity generated by wind and solar power after meeting load requirements, thus lowering the wind and solar curtailment rate. However, this also reduces the electricity costs for users, and since users consume most of their electricity during off-peak hours, the power company's revenue decreases, resulting in a decrease in overall revenue.

[0262] To analyze and compare the optimization results of the five cases, an initial electricity price of 535.72 yuan / MWh was used as the benchmark case. At this time, the load ratios of the microgrid during peak, flat, and valley periods were 31.2%, 42.8%, and 26%, respectively.

[0263] For load users, lower electricity prices result in lower electricity costs for the same load. In Case 1 and Case 2, the revenue weighting coefficient for load users is higher than that for the power company. Therefore, in these two cases, the optimal electricity prices during off-peak and peak periods are 107.1 yuan / MWh and 428.6 yuan / MWh, respectively, both being the lower limits of the off-peak and peak periods. From Case 1 to Case 3, the revenue weighting coefficient for load users gradually decreases. In Case 1, Case 2, and Case 3, the optimal electricity prices during peak periods are 792.9 yuan / MWh, 963.9 yuan / MWh, and 1085.3 yuan / MWh, respectively. In Case 4 and Case 5, the power company's weighting coefficient is higher than that for load users. Simulation results for Case 4 and Case 5 show that the revenue for load users is 2,012,191.8 yuan and 3,272,217.3 yuan, respectively, indicating that after implementing optimized electricity prices in Case 4 and Case 5, the electricity cost for users increases compared to Case 1, Case 2, and Case 3.

[0264] As shown in Tables 5 and 6, for power companies, in Case 4 and Case 5, the weighting coefficient of power company revenue is greater than that of load users. The peak-hour electricity prices in Case 4 and Case 5 are 485.5 yuan / MWh and 628.3 yuan / MWh, respectively, which are close to the upper limit of the peak-hour electricity price compared to the first three cases. The off-peak electricity price in Case 5 is 200.1 yuan / MWh, significantly higher than the off-peak prices in the first four cases. This is mainly because the power company revenue weighting in Case 5 is larger than in the other four cases.

[0265] In Case 3, the weighting factor for load user revenue and power company revenue is 0.5. Compared with the results of Cases 1 and 2, the peak-hour electricity price is higher in Case 3 because the weighting factor for power company revenue is higher than in Cases 1 and 2.

[0266] To minimize wind and solar curtailment, surplus wind and solar power output should be reduced. Surplus wind, solar, and solar power output typically occurs during off-peak and off-peak periods. Therefore, it's necessary to lower electricity prices to increase user electricity consumption, thereby reducing excess wind and solar power output.

[0267] As shown in Table 6, compared with the baseline scenario data, the off-peak load of Case 1, Case 2, Case 3, Case 4, and Case 5 increased by 2120MW, 2120MW, 2120MW, and 1649MW, respectively. This is because the off-peak electricity price for the first four cases was the lower limit of 107.1 yuan / MWh, while the off-peak electricity price for Case 5 was relatively higher. This resulted in Case 5 having the highest wind and solar curtailment rate.

[0268] The simulation results above show that as the weighting coefficient of load user revenue decreases, the electricity price during off-peak hours is the highest in Case 5, and the electricity price during normal hours increases accordingly, resulting in a certain reduction in load during normal hours and an increase in the curtailment rate of solar and wind power.

[0269] In summary, this invention provides a time-of-use pricing optimization method and terminal based on scenario sets and electricity price elasticity. Through scenario generation and scenario reduction technologies, it utilizes the probability distribution of scenario sets to reflect the randomness of wind and solar power and load. Furthermore, it proposes a time-of-use pricing optimization model that comprehensively minimizes wind and solar curtailment rates and maximizes the revenue of power companies and load users, effectively achieving time-of-use pricing optimization and providing strong support for the economic dispatch of the power system.

[0270] By constructing a scenario set, the randomness of wind power and solar power output and load is solved; by using the load price elasticity of a linear function, an optimized time-of-use pricing model is established, which can simultaneously satisfy the interests of power companies, load users, and minimize wind and solar curtailment rates.

[0271] The effectiveness of the proposed microgrid time-of-use pricing optimization method is verified through simulation examples. The results show that, compared to the baseline scenario, the proposed method can balance the economic benefits for power companies and load users while reducing wind and solar curtailment rates. The first four cases demonstrate relatively good economic performance, and all show a reduction in wind and solar curtailment rates compared to the baseline scenario.

[0272] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent modifications made based on the content of the present invention specification and drawings, or direct or indirect applications in related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A time-of-use electricity pricing optimization method based on scenario sets and electricity price elasticity, characterized in that, Including the following steps: S1. Based on the relationship between electricity price and load, construct an electricity price elasticity model; S2. Construct a scene set based on wind and solar power output and load through scene generation and scene reduction technologies; S3. Under the scenario set, construct an objective function based on the electricity price elasticity model, with the goal of minimizing the wind and solar curtailment rate and maximizing the revenue of power companies and load users, and construct a time-of-use electricity price optimization model. S4. Solve the time-of-use electricity pricing optimization model and optimize the time-of-use electricity pricing based on the solution results; Step S1 is as follows: Based on the relationship curve between electricity price and user load, the initial definition of the electricity price elasticity of load is: ； in, L 0( t (time) t The initial load, P 0( t (time) t The initial electricity price, Δ P ( t (time) t The change in electricity price, Δ L ( t (time) t The amount of load change; The electricity price elasticity model using a linear function is expressed as follows: ； in, a , b The coefficients of the linear function of load and electricity price; The price elasticity of the load can then be calculated as follows: ； Based on the impact of off-peak electricity prices in time-of-use pricing on load during flat and peak periods, the electricity price elasticity matrix is ​​expressed as follows: ； in, E P-O , E P-V , E O-V E represents the cross-elasticity coefficients for peak and off-peak periods, peak and trough periods, and off-peak and trough periods, respectively. P-P E V-V E O-O These are the self-elasticity coefficients for peak, valley, and flat periods, respectively, calculated using the formula for load price elasticity. Load after electricity price adjustment L 1 is represented as: ； in, L P,0 , L O,0 , L V,0 These represent the initial loads for peak, average, and valley periods, respectively. L 0=[ L P,0 L O,0 L V,0 ] represents the load under the initial electricity price, Δ L P Δ L O Δ L V These represent the load changes during peak, average, and off-peak periods, respectively. Step S2 includes the following steps: S21. Select and construct typical day's wind and solar power output and load data; S22. Based on the wind and solar power output and load data of the typical day, a first preset number of load and wind and solar power output scenarios were randomly generated using the super Latin cube sampling technique. S23. Using scene reduction technology, determine a second preset number of load and wind and solar power output scenes from the first preset number of load and wind and solar power output scenes to obtain a scene set.

2. The time-of-use electricity pricing optimization method based on scenario sets and electricity price elasticity according to claim 1, characterized in that, Step S21 specifically involves selecting and constructing typical day wind and solar power output and load data using the K-means clustering algorithm.

3. The time-of-use electricity pricing optimization method based on scenario sets and electricity price elasticity according to claim 2, characterized in that, Step S21 includes the following steps: S211. Obtain the M-day load and wind and solar power output data generated by the sampling; S212. Initialize the load and wind and solar power output data, and randomly extract k days of load and wind and solar power output data as the cluster center of various types of data, and set the maximum number of cycles and the change error of the cluster center; S213. Calculate the Euclidean distance between the daily sample data and the k cluster centers in the load data and wind and solar power output data. S214. Classify each of the said cluster centers into a class, and classify each of the said sample data according to the criterion of minimum Euclidean distance; S215. Recalculate the cluster centers of the data in each class; S216. Determine whether the cluster centers have changed and whether the maximum number of iterations has been reached. If the cluster centers have not changed or the maximum number of iterations has been reached, obtain the clustering result and proceed to step S217; otherwise, return to step S213. S217. Based on the comparison of the cluster center distances of the clustering results, obtain the typical scenarios of the corresponding data.

4. The time-of-use electricity pricing optimization method based on scenario sets and electricity price elasticity according to claim 3, characterized in that, The initialization of the load and wind / solar power output data in step S212 specifically involves: The load and wind / solar output data to be classified are standardized using z-score standardization: ； in, For the first m The sky t Standardized data values ​​at time points Let be the mean of the total data at time t. Let be the standard deviation of the population data at time t.

5. The time-of-use electricity pricing optimization method based on scenario sets and electricity price elasticity according to claim 1, characterized in that, The time-of-use pricing optimization model considers the revenue of power companies and load users. The objective function with the goal of maximizing is expressed as: ； Among them, Rcom l In order to be in l The revenue of the power company in this scenario, Ruser l In order to be in l In this scenario, the revenue for load users is denoted by α and β, which are the weighting coefficients for the revenue of the power company and the load users, respectively. p l for l The probability of a scene occurring is related to α and β by the formula α + β = 1. The revenue of a power company, Rcom, includes the revenue from selling electricity to customers and the cost of purchasing electricity from the main grid, expressed as: ； in, P ( t )for t Electricity price at any time L 1( t )for t The load of time, a buy for( t )yes t The electricity price purchased from the main power grid at all times d buy ( t )yes t The amount of electricity purchased from the main power grid at all times; The revenue (Ruser) of a load user is represented as: ； in, C 0 represents the electricity cost for load users under the initial electricity price. C TOU This is the electricity cost for load users calculated based on time-of-use pricing. P 0( t (time) t The initial electricity price, L 0( t (time) t The initial load, L 1( t () represents the load after adjusting the electricity price.

6. The time-of-use electricity pricing optimization method based on scenario sets and electricity price elasticity according to claim 1, characterized in that, The time-of-use electricity pricing optimization model uses the curtailment rate of solar and wind power as the key factor. The objective function with the goal of minimizing is expressed as: ； in, P char ( t ) for ESS t Charging power at any time P PV.upper ( t )and P wind.upper ( t (representing time) t The maximum value of photovoltaic power and the maximum value of wind power.

7. The time-of-use electricity pricing optimization method based on scenario sets and electricity price elasticity according to claim 1, characterized in that, The constraints in the time-of-use electricity pricing optimization model include: Power balance constraints: The power generation of wind and solar power, the charging and discharging power of energy storage systems, and the load need to meet the power balance requirements, expressed as: ； in, L 1( t This refers to the load after adjusting for electricity prices. P char ( t ) for ESS t Charging power at any time P PV ( t (time) t Photovoltaic output power, P wind ( t (time) t The wind power output power, P dis ( t ) for ESS t Discharge power at any given moment; Constraints on photovoltaic and wind power generation: The power output of wind power and photovoltaic power is mainly determined by the maximum and minimum values ​​of wind power and photovoltaic power, respectively, as follows: ； ； in, P PV.upper ( t (time) t The maximum photovoltaic power, P wind.upper ( t (time) t The maximum wind power output; ESS constraints: During the charging process of the ESS, the battery pack is powered by surplus wind and solar power. The stored energy of the battery is determined by the charging power and the stored energy in the previous hour, as follows: ； in, E ( t )for t The energy stored in the battery at all times. For the charging efficiency of the storage battery; When the ESS is discharging, the battery pack supplies power to the load. The stored energy of the battery is determined by the discharge power and the stored energy of the previous hour, expressed as: ； in, P dis This refers to the battery's discharge power. The discharge efficiency of the battery; SOC constraints are expressed as follows: ； in, S OC ( t )for t ESS's SOC at all times S OC.lower This represents the minimum state of charge of the ESS. S OC.upper This indicates the maximum state of charge of the ESS; The discharge power constraint of the ESS is expressed as: ； The charging power constraint of the ESS is expressed as follows: ； in, P char.upper and P dis.upper These are the maximum charging power and the maximum discharging power of the ESS, respectively. ESS This refers to the rated capacity of the battery pack.

8. A time-of-use electricity pricing optimization terminal based on scenario sets and electricity price elasticity, comprising a processor, a memory, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps in the time-of-use electricity pricing optimization method based on scenario sets and electricity price elasticity as described in any one of claims 1-7.

Images