Power distribution network flexibility demand determination method considering extreme climate influence

Abstract

The invention relates to a power distribution network flexibility demand determination method considering extreme climate influence, and belongs to the technical field of power system dispatching planning. The method comprises the following steps: determining a plurality of flexibility demand typical scenes and corresponding scene probabilities under the extreme climate based on an extreme climate element data set, and constructing a flexibility demand uncertainty fuzzy set based on the scene probabilities; obtaining the output of a wind turbine generator, the output of a photovoltaic unit, a basic load and a climate sensitive load under the extreme climate based on the typical scene of the flexibility demand; constructing a distribution robust optimization model taking the minimum operation cost of the regional power distribution network as a target; an original target is decomposed into a main problem and a sub-problem, and Camp is used; and carrying out iterative solution on the model to obtain an operation scheme of each unit of the regional power distribution network and a flexibility requirement for adapting to extreme climate. According to the method, the multi-scale flexibility demand quantification of the power distribution network adapting to the extreme climate is accurately realized.

Description

Technical Field

[0001] The present invention relates to the technical field of power system dispatching planning, and particularly to a method for determining the flexibility requirements of a distribution network considering the impact of extreme climate. Background Art

[0002] The flexibility requirements of the distribution network mainly stem from the uncertainty and volatility of load and new energy. The greater the fluctuation of the net load (the power difference between load and new energy), the greater the flexibility requirements. Currently, there are mainly two methods for analyzing the flexibility requirements of the distribution network: the first is through quantification and evaluation methods to quantitatively analyze the flexibility requirements in the system; the second is dynamic simulation evaluation, that is, setting the operation scenarios of the power system for optimal dispatching and production simulation to analyze the flexibility requirements of the power system. However, extreme climates such as high-temperature drought, low-temperature cold snap, continuous windless and lightless, and strong wind and strong light will have uncertain impacts on new energy power generation and load demand, further affecting the fluctuation of the net load curve. Existing methods do not consider the non-linear coupling impact of extreme climate on the fluctuation of new energy output and load mutation, resulting in a deviation between the flexibility resource allocation and the actual demand, and cannot meet the operation requirements of the power system under extreme climate. Summary of the Invention

[0003] In view of the above analysis, the present invention aims to disclose a method for determining the flexibility requirements of a distribution network considering the impact of extreme climate. By analyzing the impact of extreme climate on new energy power generation and load, constructing an uncertain fuzzy set of flexibility requirements in extreme climate situations, and evaluating the regulation ability of existing flexibility resources in the distribution network, the multi-scale flexibility requirements of the distribution network can be quantified.

[0004] A method for determining the flexibility requirements of a distribution network considering the impact of extreme climate provided by the present invention specifically includes the following steps: Based on the extreme climate element data set, determine multiple typical scenarios of flexibility requirements and corresponding scenario probabilities under extreme climate, and construct an uncertain fuzzy set of flexibility requirements based on the scenario probabilities; Based on the typical scenarios of flexibility requirements, obtain the output of wind turbines, the output of photovoltaic units, the basic load, and the climate-sensitive load under extreme climate; Construct a distributionally robust optimization model with the goal of minimizing the operating cost of the regional distribution network; by decomposing the original goal into a master problem and a sub-problem, using the C&CG algorithm, and based on the output of wind turbines, the output of photovoltaic units, the basic load, the climate-sensitive load, the set deterministic constraints, and the uncertain fuzzy set, iteratively solve the model to obtain the operating schemes of each unit in the regional distribution network and the flexibility requirements for adapting to extreme climate.

[0005] Further, determining the scenario probabilities corresponding to multiple typical scenarios of flexibility requirements under extreme climate based on the extreme climate element dataset, and constructing an uncertain fuzzy set of flexibility requirements based on the scenario probabilities includes: Using the K-means clustering algorithm to cluster based on the extreme climate element dataset to obtain multiple typical scenarios of flexibility requirements under extreme climate and the corresponding scenario probabilities; Constructing a confidence set based on the 1-norm and ∞-norm based on the scenario probabilities; Constructing an uncertain fuzzy set of flexibility requirements based on the confidence set of the 1-norm and ∞-norm.

[0006] Further, obtaining the wind power, photovoltaic unit output, base load, and climate-sensitive load under extreme climate based on the typical scenarios of flexibility requirements includes: Based on the climate data of the typical scenarios of flexibility requirements, using the wind turbine output model and the photovoltaic unit output model to calculate the wind power and photovoltaic unit output respectively; Based on the climate data of the typical scenarios of flexibility requirements, using the trained hybrid model of CNN and LSTM to predict the corresponding base load and climate-sensitive load.

[0007] Further, taking the minimum operation cost of the regional distribution network as the objective is expressed by the objective function as: ; ; where is the power purchase cost of the regional distribution network from the superior power grid at time , , are the power generation costs of the gas, wind power, and photovoltaic units of the regional distribution network at time respectively; is the scheduling cost of the energy storage unit of the regional distribution network at time ; is the cost of implementing price-based demand response of the regional distribution network; is the incentive-based demand response compensation cost; is the interaction power between the regional distribution network and the superior power grid at time and are the load demands after and before implementing price-based demand response at time respectively; is the electricity price change value at time is the power purchase cost of the regional distribution network from the superior power grid at time; the superscript and Indicates the unit type; Is a function between the unit operation cost and the output; Is The unit output at time , , , Respectively represent gas, energy storage, wind power and photovoltaic units; , Are the start-up and shutdown states of the unit; , Are the start-up and shutdown costs of the unit; Is the unit interruptible load compensation cost; Is the interruptible load at The response power at time

[0008] Furthermore, the flexible demand uncertain fuzzy set is expressed as: ; Wherein, Is the flexible demand uncertain fuzzy set; Is the typical scenario of flexible demand The set of positive real numbers of the probability distribution; Is the typical scenario of flexible demand The true occurrence probability of; The number of typical scenarios of flexible demand; Is the occurrence probability of the typical scenario obtained by clustering; Is the total number of samples in the extreme climate element data set; , Are the confidence levels satisfied by the scenario probability distributions based on the 1-norm and -norm respectively.

[0009] Furthermore, the training process of the trained CNN and LSTM hybrid model includes; Determine the historical reference load and historical climate-sensitive load at the corresponding time based on the total load at the corresponding time of the historical climate element data; Train the CNN and LSTM hybrid model based on the historical climate element data, historical reference load and historical climate-sensitive load to form a training set, and obtain the trained CNN and LSTM hybrid model.

[0010] Furthermore, the set deterministic constraints include power balance constraints, and the power balance constraints are expressed as: ; Wherein, Is The interactive power between the regional distribution network and the superior power grid at time , , are respectively the output of the gas, wind power and photovoltaic units at time is the load demand after implementing price-based demand response at time is the interruptible load at the response power at time , are respectively the charge and discharge power of the energy storage at time

[0011] Furthermore, using the price demand response constraint, based on the load demand before implementing price-based demand response, determine the load demand after implementing price-based demand response at time , and the price demand response constraint is expressed as: ; wherein, , , are respectively the electricity prices during peak hours, valley hours and flat hours; , are respectively the lowest and highest electricity prices; is the peak-valley electricity price ratio; , are respectively the minimum and maximum peak-valley electricity price ratios; is the number of scheduling periods.

[0012] Furthermore, based on the reference load and the climate-sensitive load, determine the load demand before implementing price-based demand response, including: ; wherein, is the reference load predicted based on historical reference load data; is the climate-sensitive load.

[0013] Furthermore, the historical reference load is expressed as: ; wherein, is the historical reference load; represents the load represented by the th curve among the five curves with the lowest average historical weekday reference load; represents the load represented by the th curve among the five curves with the lowest average historical weekend reference load.

[0014] The present invention can at least achieve one of the following beneficial effects: By characterizing the variation characteristics of new energy output and load demand under extreme climates, constructing a typical scenario set for flexibility demand, and evaluating the regulation capabilities of existing flexibility resources in the distribution network, the multi-scale flexibility demand quantification evaluation of the distribution network is realized, solving the problem that there is a deviation between the flexibility resource allocation and the actual demand in traditional distribution network planning due to the failure to consider the non-linear coupling effect of extreme climates on new energy output fluctuations and load mutations, and improving the new energy consumption rate and the reliability of power grid supply guarantee.

[0015] By constructing a training set based on historical climate element data and corresponding climate-sensitive loads, for the CNN and LSTM hybrid model, the accuracy of the model for predicting climate-sensitive loads is improved; by performing cluster analysis on the extreme climate element data set to obtain a typical scenario set for flexibility demand, the impact of extreme climates on new energy output and load demand is accurately characterized; by modeling the demand-side response, the potential of demand-side load response in actual application scenarios is accurately described. The accuracy of determining the flexibility demand of the distribution network under extreme climates is improved, which can provide a scientific basis for the flexibility resource allocation of the power system.

[0016] Other features and advantages of the present invention will be described in the subsequent description, and some advantages can be made obvious from the description, or understood by implementing the present invention. The objectives and other advantages of the present invention can be achieved and obtained through the content specifically pointed out in this application document. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The drawings are only for the purpose of showing specific embodiments and are not considered as limiting the present invention. Throughout the drawings, the same reference signs represent the same components; Figure 1 is a flowchart of the present invention; Figure 2 is a diagram for generating typical scenarios of flexibility demand in an embodiment of the present invention; Figure 3 is a price demand curve in an embodiment of the present invention; Figure 4 is a user response curve in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] The following will specifically describe the preferred embodiments of the present invention in conjunction with the drawings, where the drawings form a part of this application and are used together with the embodiments of the present invention to explain the principles of the present invention, and are not used to limit the scope of the present invention.

[0019] An embodiment of the present invention discloses a method for determining the flexibility demand of a distribution network considering the impact of extreme climates, as Figure 1 shown, specifically including steps S1-S3.

[0020] Specifically, the extreme climates include sudden climate events such as high temperature, heavy rain, cold snap, typhoon, etc.

[0021] Specifically, during the operation of the distribution network, the difference between the electricity consumption load and the new energy power causes the flexibility demand of the distribution network. When extreme climate events occur, it will have an uncertain impact on new energy power generation and electricity consumption load, resulting in a large fluctuation in flexibility demand. Therefore, by analyzing typical scenarios of extreme climates, typical scenarios of the flexibility demand of the distribution network can be obtained.

[0022] Step S1: Determine multiple typical scenarios of flexibility demand and corresponding scenario probabilities under extreme climates based on the historical extreme climate element dataset, and construct an uncertain fuzzy set of flexibility demand based on the scenario probabilities. It includes: S11: Use the K-means clustering algorithm to cluster based on the extreme climate element dataset to obtain multiple typical scenarios of flexibility demand and corresponding scenario probabilities under extreme climates.

[0023] Specifically, before clustering, based on the historical climate element data and the climate-sensitive load at the corresponding time, screen the climate elements that have a greater impact on the climate-sensitive load. Use Pearson correlation to measure the correlation degree between the climate-sensitive load and the climate element variables, and determine the climate elements with a greater correlation degree as the clustering basis. Exemplarily, the climate elements with a greater correlation degree obtained through calculation include temperature, wind speed, light intensity, humidity, etc.

[0024] Specifically, the clustering process is as Figure 2 shown, including: a. Set the number of clusters to , and randomly select the initial clustering centers of each climate element for each category; b. Calculate the Euclidean distances between each climate element sample and the clustering centers, and assign the sample to the class where the clustering center with the smallest distance is located; c. Update the clustering center of each category to the average value of all samples in that category d. Repeat steps b and c until all clustering centers do not change or reach the maximum number of iterations.

[0025] Finally, multiple typical scenarios of flexibility demand under extreme climates are obtained; the scenario probabilities of each typical scenario of flexibility demand are: ; where is the number of samples in the th class; is the total number of samples.

[0026] S12: Construct a confidence set based on the 1-norm and ∞-norm based on the scenario probabilities.

[0027] Specifically expressed as: ; ; Among them, and are the confidence sets of the 1-norm and ∞-norm respectively; is the true distribution of the scenario occurrence probability; is the probability distribution of the typical scenarios obtained by clustering; is the th element in; is the set of typical scenarios; is the occurrence probability of the typical scenario; and are the allowable limit values of the probability deviation under the 1-norm and -norm constraints respectively.

[0028] S13. Construct a flexible demand uncertainty fuzzy set based on the confidence sets of the 1-norm and ∞-norm.

[0029] Specifically, the flexible demand uncertainty fuzzy set is expressed as: ; Among them, is the flexible demand uncertainty fuzzy set; is the set of positive real numbers of the probability distribution of the typical scenarios of the flexible demand ; is the true occurrence probability of the typical scenario of the flexible demand ; the number of typical scenarios of the flexible demand; is the occurrence probability of the typical scenario obtained by clustering; is the total number of samples in the extreme climate element data set; , are the confidence levels satisfied by the scenario probability distribution based on the 1-norm and -norm respectively.

[0030] Step S2. Obtain the wind turbine output, photovoltaic unit output, and climate-sensitive load under extreme climate based on the typical scenarios of the flexible demand.

[0031] S21. Based on the climate element data of the typical scenarios of the flexible demand, use the wind turbine output model and the photovoltaic unit output model to calculate the wind power and photovoltaic unit output respectively.

[0032] Specifically, the wind turbine output model is expressed as: ; Among them, is the real-time output of the wind turbine at moment; is the ambient air density where the wind turbine is located; is the blade radius of the wind turbine; is the efficiency coefficient of the wind turbine; is the environment where the wind turbine is located wind speed at moment.

[0033] Furthermore, to ensure the safety and stability of the wind turbine, the real-time output is restricted by the cut-in wind speed , cut-out wind speed and rated wind speed : .

[0034] Furthermore, the air density will decrease with the increase of temperature and is a variable value. Its calculation method is: ; Among them, is the molar mass of air, with the unit of kg / mol; is the ambient pressure where the wind turbine is located, with the unit of Pa; is the ideal gas constant, with the unit of J / (mol·K); is the Celsius temperature at the moment in the environment where the wind turbine is located , with the unit of °C.

[0035] Specifically, the output model of the photovoltaic unit is expressed as: ; Among them, is the real-time output of the photovoltaic unit at moment, with the unit of MW; is light intensity at moment, unit ; is ambient temperature at moment, with the unit of °C; is the output of the photovoltaic unit under standard test conditions.

[0036] S22. Based on the climate data of the typical scenario of the flexibility demand, use the trained hybrid model of CNN and LSTM to predict the corresponding climate-sensitive load.

[0037] Specifically, the training process of the trained hybrid model of CNN and LSTM includes; S22-b. Determine the historical baseline load and historical climate-sensitive load at the corresponding time based on the total load at the corresponding time of the historical climate element data; the historical climate element data includes wind speed, temperature, light intensity, etc. S22-c. Construct a training set based on the historical climate element data, the historical baseline load, and the climate-sensitive load at the corresponding time to train the CNN and LSTM hybrid model, and obtain the trained CNN and LSTM hybrid model.

[0038] Further, in S22-b, the climate-sensitive load at the corresponding time of the historical climate element data is the difference between the load curve and the historical baseline load curve: ; where is the climate-sensitive load at the corresponding time of the historical climate element data, reflecting the impact of climate elements on the load; is the historical baseline load, reflecting the overall development trend of the load over a long period of time; is the historical total load at the moment; Further, the historical baseline load is expressed as: ; where is the historical baseline load; represents the load represented by the th curve among the five curves with the lowest average baseline load on historical working days; represents the load represented by the th curve among the five curves with the lowest average baseline load on historical weekends. Exemplarily, the following introduces the method for determining the historical baseline load: Select the working day and weekend baseline load curves with the daily average temperature between 10°C and 24°C in spring (March - May) and autumn (September - November) from the historical baseline load data as the alternative baseline load curves, and calculate the historical baseline load based on the 5 curves with the lowest average baseline load on working days and weekends respectively selected from the alternative load curves.

[0039] Further, when training the hybrid model in S22-c, use the mean absolute percentage error root mean square error and prediction accuracy three indicators to evaluate the prediction performance of the hybrid model.

[0040] Step S3: Construct a distributionally robust optimization model with the objective of minimizing the operating cost of the regional distribution network; by decomposing the original objective into a master problem and a sub-problem, using the C&CG algorithm, and based on the output of the wind turbine units, the output of the photovoltaic units, the base load, the climate-sensitive load, the set deterministic constraints, and the uncertain fuzzy set, iteratively solve the model to obtain the operating plans of each unit of the regional distribution network and the flexibility requirements for adapting to extreme climates.

[0041] Specifically, in step S3, first, model various flexibility resources such as gas (the gas turbine unit in this invention is a gas turbine), energy storage, and the demand side to describe the operating constraints and flexibility regulation capabilities of various flexibility resources.

[0042] Specifically, model the gas turbine. The upward and downward flexibility that the gas turbine can provide is expressed as: ; ; where and are the upward and downward flexibility that the gas turbine can provide at time respectively; is the output of the gas turbine at time ; , are the maximum and minimum outputs of the gas turbine; , are the upward and downward ramp rates of the gas turbine.

[0043] Specifically, model the energy storage unit, including the charge and discharge power model of the energy storage unit and the upward and downward flexibility model of the energy storage unit.

[0044] The charge and discharge power model of the energy storage unit is expressed as: ; where represents the charge and discharge power of the energy storage unit at time ; and are the charge and discharge powers of the energy storage unit at time respectively; , are auxiliary binary variables representing the charge and discharge states of the energy storage unit at time ; is the state of charge of the energy storage unit at time ; and are the charge and discharge efficiencies of the energy storage unit respectively.

[0045] The flexibility models for upward and downward regulation of the energy storage unit are expressed as: ; ; where and are the upward and downward regulation flexibilities that the energy storage unit can provide at time respectively; is the charge-discharge power of the energy storage unit at time ; , are the charging and discharging powers of the energy storage unit at time respectively; , are the charge and discharge state variables of the energy storage unit; is the state of charge of the energy storage unit at time ; and are the charging and discharging efficiencies of the energy storage respectively; and are the rated charging and discharging powers of the energy storage respectively; and are the minimum charging and discharging powers of the energy storage; and are the maximum and minimum energy storage capacities.

[0046] Specifically, the price-based demand response on the demand side is modeled and expressed as: ; where is the load demand after implementing the price demand response at time ; is the price elasticity matrix, , is the number of scheduling periods. The elements on the diagonal are the self-elasticity coefficients, which describe the impact of the relative change in electricity price in this period on the electricity demand in this period, and the remaining elements are the cross-elasticity coefficients, which describe the impact of the relative change in electricity price in this period on the electricity demand in the remaining periods. Among them, the self-elasticity coefficient should be negative, and the cross-elasticity coefficient should be positive; is the price elasticity coefficient ( ), , when , it is called the self-elasticity coefficient, and when , it is called the cross-elasticity coefficient (The price elasticity coefficient explains the impact of electricity price changes on the changes in electricity demand, reflecting the relationship between the change rate of electricity demand and the change rate of electricity price, as shown in Figure 3 ); , are the load demand and the change value of load demand at time respectively; , are respectively the initial electricity price at a moment and the electricity price change value; Specifically, model the incentive-based demand response on the demand side. The incentive-based demand response reaches an agreement with users through economic compensation or reward contracts to adjust the electricity load. Among them, interruptible load is the most common and widely used form. During peak load or emergencies, an interruption application is sent to users participating in the demand response. After receiving the dispatching instruction, the users actively reduce part of their electricity load, thereby reducing the system operation risk and obtaining economic compensation; without the dispatching instruction, the users' load maintains normal operation.

[0047] Furthermore, the calculation method of the compensation cost for the incentive-based demand response modeling is: ; wherein, is the compensation cost per unit of interruptible load; is the interruptible load at the response power at a moment.

[0048] Furthermore, after modeling the price-based demand response and the incentive-based demand response on the demand side, a demand response model considering user responsiveness is constructed, including a price-based demand response model considering user responsiveness and an incentive-based demand response model considering user responsiveness.

[0049] It should be noted that since the price-based demand response model and the incentive-based demand response model on the demand side are constructed based on the ideal condition that users can use electricity rationally. However, in reality, the demand response behavior of users is also affected by factors such as the publicity intensity of demand response policies and the behavior of user groups. The relationship between the participation degree of demand response and the information intensity is as Figure 4 shown. In the figure, the horizontal axis is the information intensity, and the vertical axis is the user responsiveness. The user responsiveness curve has a dead zone, a linear zone, and a saturation zone. When in the dead zone, the user demand is not obvious to the electricity price change, and the demand response responsiveness is 0, which is because the economic benefits brought by users' participation in the demand response are not high; when in the linear zone, the user responsiveness increases with the electricity price change; when in the saturation zone, the user responsiveness reaches the maximum value and remains unchanged with the increase of the electricity price information intensity.

[0050] Therefore, the price demand response model considering user responsiveness is constructed as: ; Converted to: ; wherein, is the user responsiveness of the price-based demand response considering time-of-use electricity price.

[0051] Furthermore, the incentive-based demand response model considering user responsiveness is constructed as follows: ; where is the user responsiveness of the incentive-based demand response considering interruptible load; is the maximum response power of the interruptible load at time.

[0052] Specifically, in step S3, the objective function with the minimum operating cost of the regional distribution network is expressed as: ; ; where is the power purchase cost of the regional distribution network from the superior grid at time; , , are the power generation costs of the gas, wind power, and photovoltaic units of the regional distribution network at time, respectively; is the scheduling cost of the energy storage unit of the regional distribution network at time; is the cost of implementing price-based demand response in the regional distribution network; is the compensation cost for incentive-based demand response; is the interactive power between the regional distribution network and the superior grid at time; and are the load demands after and before implementing price-based demand response at time, respectively; is the electricity price change value at time; is the power purchase cost of the regional distribution network from the superior grid at time; the superscripts and represent the unit types; is the function between the unit operating cost and the output; is the unit output at time; , , , represent gas, energy storage, wind power, and photovoltaic units, respectively; , are the start-up and shut-down states of the unit; , are the start-up and shut-down costs of the unit; is the unit interruptible load compensation cost; is the response power of the interruptible load at moment.

[0053] Specifically, in step S3, the set deterministic constraints include gas turbine output constraints, energy storage constraints, power balance constraints, and price-based demand response constraints.

[0054] Furthermore, the gas turbine output constraint is expressed as: ; where is the operating state of the gas turbine at moment; , are the start-up and shut-down states of the gas turbine at moment; , are the minimum continuous start-up and shut-down times of the gas turbine respectively; is the output of the gas turbine at moment; , are the minimum and maximum outputs of the gas turbine respectively; , are the maximum allowable upward and downward ramp rates of the gas turbine respectively; is the time interval at each moment.

[0055] Furthermore, the energy storage constraint is expressed as: ; where is the charge and discharge power of the energy storage unit at moment; , are the charging and discharging powers of the energy storage unit at moment respectively; , are the charge and discharge state variables of the energy storage unit; is the state of charge of the energy storage unit at moment; and are the charging and discharging efficiencies of the energy storage respectively; and are the rated charging and discharging powers of the energy storage respectively; and are the minimum charging and discharging powers of the energy storage; and are the maximum and minimum energy storage capacities.

[0056] Furthermore, the power balance constraint is expressed as: ; Among them, is the interactive power between the time-area distribution network and the superior power grid; , , , are respectively the outputs of gas, energy storage, wind power and photovoltaic units at time is the load demand after implementing price-based demand response at time is the interruptible load at the response power at time is the energy storage charging power at time

[0057] Furthermore, using the price demand response constraint, based on the load demand before implementing the price-based demand response, determine the load demand after implementing the price-based demand response at time .

[0058] Specifically, the price demand response constraint is expressed as: ; Among them, , , are respectively the electricity prices during peak hours, valley hours and flat hours; , are respectively the minimum and maximum electricity prices; is the peak-valley electricity price ratio; , are respectively the minimum and maximum peak-valley electricity price ratios; is the number of scheduling periods.

[0059] Furthermore, based on the reference load and the climate-sensitive load, determine the load demand before implementing the price-based demand response, including: ; Among them, is the reference load under extreme climate predicted based on historical reference load data; is the climate-sensitive load.

[0060] Furthermore, the historical reference load has been described in step S22-b.

[0061] Specifically, in order to find the operating plans of each unit under the probability distribution of the worst-case scenario, a distributionally robust analysis model for the flexibility requirements of a distribution network is established based on the models constructed by various types of flexibility resources, the objective function, the set deterministic constraints, and the uncertain fuzzy set, and the C&CG algorithm is used to iteratively solve the distributionally robust optimization model. The distributionally robust optimization model is a two-stage three-layer optimization problem of min-max-min.

[0062] Its compact form is expressed as: ; (1) where is an integer variable, including the start-stop plan of the gas turbine and the charge / discharge state of the energy storage; is all continuous decision variables affected by the uncertain terms (wind power, photovoltaic power, load), including the output of the gas turbine, the charge / discharge power of the energy storage, and the demand response power; , is the feasible set; , are the coefficient vectors of the objective function; is the fuzzy set; , , are the constraint coefficient matrices; , are the constraint constant vectors.

[0063] Furthermore, the C&CG algorithm is used to iteratively solve the distributionally robust optimization model, and the original objective is decomposed into a master problem and a subproblem.

[0064] Master problem: Given the probability distribution of the flexibility requirement scenarios , solve the optimal solution that satisfies the system economy and provide a lower bound value for Equation (1).

[0065] ; (2) where is the set threshold; is the number of iterations.

[0066] Subproblem: The subproblem, under the condition that the master problem obtains the variable , finds the worst-case probability distribution of the typical scenarios of flexibility requirements, returns it to the master problem, and provides an upper bound value for Equation (1): ; (3) Since the outer and inner problems have the structural decoupling property in the constraint space, Equation (3) can be decomposed into two steps for solution, as shown in Equations (4) and (5): ; (4) ; (5) The iterative solution process is as follows: ① Set the upper and lower bounds of the initial objective function to be and ; The number of iterations is , and the convergence threshold ; ② Solve the master problem to obtain the optimal solution , and update the lower bound of the objective function ; ③ Based on in step ②, solve the sub-problem to obtain the optimal solution and the worst probability distribution , and update the upper bound of the objective function ; ④ Determine whether to converge. If is satisfied, stop the iteration and output , , ; Otherwise, let , and go back to step ② to continue the solution until convergence.

[0067] Finally, the operating schemes of each unit in the regional distribution network and the flexibility requirements for adapting to extreme climates are obtained. The flexibility requirements include fluctuations in wind power output, photovoltaic output, and load demand caused by extreme climates.

[0068] A method for determining the flexibility requirements of a distribution network considering the impact of extreme climates disclosed in this embodiment, by characterizing the variation characteristics of new energy output and load demand under extreme climates, constructing a typical scenario set of flexibility requirements, and evaluating the regulation ability of existing flexibility resources in the distribution network, thus realizing the quantitative evaluation of the multi-scale flexibility requirements of the distribution network, solving the problem that the deviation between the flexibility resource allocation and the actual demand exists in traditional distribution network planning due to the non-linear coupling impact of extreme climates on new energy output fluctuations and load demand, and improving the new energy consumption rate and the reliability of power grid power supply.

[0069] By constructing a training set based on historical climate element data and corresponding climate-sensitive loads, the accuracy of climate-sensitive load prediction is improved for the CNN and LSTM hybrid models; by performing clustering analysis on the extreme climate element data set, a typical scenario set of flexibility requirements is obtained, accurately depicting the impact of extreme climates on new energy output and load demand; by modeling the demand-side response, the potential of demand-side load response in actual application scenarios is accurately described. The accuracy of determining the flexibility requirements of the distribution network under extreme climates is improved, which can provide a scientific basis for the flexibility resource allocation of the power system.

[0070] As described above, it is only the preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A method for determining the flexibility requirements of a distribution network considering the impact of extreme climates, characterized in that It includes the following steps: Based on the extreme climate element dataset, determine multiple typical scenarios of flexibility requirements under extreme climate and the corresponding scenario probabilities, and construct an uncertain fuzzy set of flexibility requirements based on the scenario probabilities; Based on the typical scenarios of flexibility requirements, obtain the output of wind turbines, the output of photovoltaic units, the base load, and the climate-sensitive load under extreme climate; Construct a distributionally robust optimization model with the goal of minimizing the operating cost of the regional distribution network; by decomposing the original goal into a master problem and a subproblem, using the C&CG algorithm, and based on the output of wind turbines, the output of photovoltaic units, the base load, the climate-sensitive load, the set deterministic constraints, and the uncertain fuzzy set, iteratively solve the model to obtain the operating plans of each unit of the regional distribution network and the flexibility requirements to adapt to extreme climate.

2. The method for determining the flexibility requirements of a distribution network according to claim 1, characterized in that The determining the corresponding scenario probabilities of multiple typical scenarios of flexibility requirements under extreme climate based on the extreme climate element dataset and constructing an uncertain fuzzy set of flexibility requirements based on the scenario probabilities includes: Using the K-means clustering algorithm, cluster based on the extreme climate element dataset to obtain multiple typical scenarios of flexibility requirements under extreme climate and the corresponding scenario probabilities; Construct a confidence set based on the 1-norm and ∞-norm based on the scenario probabilities; Construct an uncertain fuzzy set of flexibility requirements based on the confidence set based on the 1-norm and ∞-norm.

3. The method for determining the flexibility requirements of a distribution network according to claim 1 or 2, characterized in that The obtaining the output of wind power, photovoltaic units, the base load, and the climate-sensitive load under extreme climate based on the typical scenarios of flexibility requirements includes: Based on the climate data of the typical scenarios of flexibility requirements, use the wind turbine output model and the photovoltaic unit output model to calculate the output of wind power and photovoltaic units respectively; Based on the climate data of the typical scenarios of flexibility requirements, use the trained CNN and LSTM hybrid model to predict the corresponding base load and climate-sensitive load.

4. The method for determining the flexibility requirements of a distribution network according to claim 2, wherein The goal of minimizing the operating cost of the regional distribution network is expressed by the objective function as: ； ； Among them, is the electricity purchase cost of the regional distribution network from the superior power grid at time ; ; are respectively the power generation costs of the gas, wind power, and photovoltaic units of the regional distribution network at time; is the scheduling cost of the energy storage unit of the regional distribution network at time; is the cost of implementing price-based demand response for the regional distribution network; is the compensation cost for incentive-based demand response; is the interactive power between the regional distribution network and the superior power grid at time and are respectively the load demands after and before implementing price-based demand response at time is the electricity price change value at time is the electricity purchase cost of the regional distribution network from the superior power grid at time ; the superscripts and represent the unit types; is the function between the unit operation cost and the output; is the unit output at time ; ; ; respectively represent gas, energy storage, wind power, and photovoltaic units; ; are the start-up and shut-down states of the unit; ; are the start-up and shut-down costs of the unit; is the unit interruptible load compensation cost; is the response power of the interruptible load at time 5. The method for determining the flexibility requirements of a distribution network according to claim 4, wherein The uncertain fuzzy set of flexibility requirements is expressed as: ； Among them, is the uncertain fuzzy set of flexibility requirements; is the typical scenario of flexibility requirements the set of positive real numbers of the probability distribution; is the typical scenario of flexibility requirements the true occurrence probability; the number of typical scenarios of flexibility requirements; is the occurrence probability of the typical scenario obtained by clustering; is the total number of samples in the extreme climate element dataset; 、 are the confidence levels satisfied by the scenario probability distributions based on the 1-norm and -norm respectively.

6. The method for determining the flexibility requirements of a distribution network according to claim 3, characterized in that The training process of the trained CNN and LSTM hybrid model includes; Based on the total load at the corresponding time of the historical climate element data, determine the historical reference load and the historical climate-sensitive load at the corresponding time; Based on the historical climate element data, the historical reference load, and the historical climate-sensitive load, construct a training set to train the CNN and LSTM hybrid model to obtain the trained CNN and LSTM hybrid model.

7. The method for determining the flexibility requirements of a distribution network according to claim 4, characterized in that The set deterministic constraints include a power balance constraint, and the power balance constraint is expressed as: ； Among them, is the interactive power between the regional distribution network and the superior power grid at time , , are respectively the output powers of gas, wind power and photovoltaic units at time is the load demand after implementing price-based demand response at time is the response power of the interruptible load at time , are respectively the charge and discharge powers of the energy storage at time 8. The method for determining the flexibility requirements of a distribution network according to claim 7, wherein Using price-demand response constraints, based on the load demand before implementing price-based demand response , determine the load demand after implementing price-based demand response at the moment , and the price-demand response constraint is expressed as: ； Among them, , , are the electricity prices during peak hours, valley hours and flat hours respectively; , are the lowest and highest electricity prices respectively; is the peak-valley electricity price ratio; , are the minimum and maximum peak-valley electricity price ratios respectively; is the number of scheduling periods.

9. The method for determining the flexibility requirements of a distribution network according to claim 8, wherein Determine the load demand before implementing the price-based demand response based on the benchmark load and the climate-sensitive load , including: ； Among them, is the baseline load predicted based on historical baseline load data; is the climate-sensitive load.

10. The method for determining the flexibility requirements of a distribution network according to claim 9, characterized in that, The historical reference load is expressed as: ； Among them, is the historical baseline load; represents the load represented by the th curve among the five curves with the lowest average historical weekday baseline load; represents the load represented by the th curve among the five curves with the lowest average historical weekend baseline load.

Images