ADN source-load-storage collaborative double-layer optimization configuration method considering demand response

Abstract

The invention discloses an ADN source-load-storage collaborative double-layer optimization configuration method considering demand response, and the method comprises the following steps: building an ADN source-load-storage model based on the demand response, which comprises a planning layer and an operation layer; a Metropolis standard in an SA algorithm is fused in a quantum behavior particle swarm optimization algorithm, and a dynamic adjustment strategy is introduced in Levy flight of a cuckoo search algorithm; and performing algorithm fusion on the improved quantum behavior particle swarm optimization algorithm and the cuckoo search algorithm to obtain an improved hybrid cuckoo search quantum behavior particle swarm optimization algorithm, and solving variables in the planning layer and the operation layer to complete configuration. According to the method, the double-layer planning model is established, the coordination of planning and operation is realized, and the reliability, the safety and the sustainability of the power distribution network in the operation process are improved while the investment and operation cost of the active power distribution network is reduced.

Description

Technical Field

[0001] The present invention belongs to the technical field related to active distribution network optimization, and in particular relates to an ADN source-load-storage collaborative double-layer optimization configuration method considering demand response. Background Art

[0002] At present, the world is facing a series of severe challenges such as increasingly severe energy resource constraints, prominent ecological and environmental problems, and climate change. Large-scale access to distributed power sources is the only way forward. However, the output of wind and photovoltaic power generation has strong randomness and timing characteristics. When these power sources are connected to the distribution network in large quantities, they may cause problems such as grid voltage fluctuations and resource waste. In order to meet these challenges, the power grid needs to have higher scheduling flexibility, management capabilities, and advanced intelligent and automation technologies. Therefore, the concept of active distribution network (ADN) came into being. Its core concept is to achieve two-way interaction and coordinated optimization between the power grid and various energy resources through information and automation technology. ADN can respond to changes in the power grid and energy market in real time, optimize resource allocation, and thereby improve the operating efficiency, reliability and economy of the power grid, and better adapt to the volatility and uncertainty of renewable energy.

[0003] For active distribution networks with a high proportion of new energy access, reasonable site selection and capacity planning for distributed power sources can not only reduce the investment and operation costs of wind power, photovoltaic and other equipment, but also effectively avoid problems such as low energy utilization efficiency, large system voltage deviation and load fluctuation caused by excessive concentration of distributed power sources. Demand response, as the core means of source-load interaction, is an effective strategy for distribution networks to dispatch flexible load resources on the demand side.

[0004] Therefore, it is urgent to propose an ADN source-load-storage collaborative two-layer optimization configuration method that considers demand response. Summary of the invention

[0005] To solve the above technical problems, the present invention proposes a two-layer optimization configuration method for ADN source-load-storage collaboration considering demand response. The upper layer of the model is a planning layer, which is used to obtain the optimal location and capacity solution for each device in the active distribution network; the lower layer is an operation layer, which is used to obtain the optimal scheduling solution for each flexible resource including the demand response load; in order to solve the complexity of the two-layer planning problem in the solution process, an improved hybrid cuckoo search quantum behavior particle swarm algorithm is adopted to increase the computational efficiency and avoid falling into the local optimal solution, thereby solving the possible misalignment problem between "planning" and "operation" in the optimization of the active distribution network, so as to solve the problems existing in the above-mentioned prior art.

[0006] To achieve the above object, the present invention provides an ADN source-load-storage collaborative dual-layer optimization configuration method considering demand response, comprising the following steps:

[0007] Building an ADN source-load-storage model based on demand response, wherein the ADN source-load-storage model includes a planning layer and an operation layer;

[0008] The Metropolis criterion in the SA algorithm is integrated into the quantum-behaved particle swarm optimization algorithm, and a dynamic adjustment strategy based on the Levy distribution is introduced in the Levy flight of the cuckoo search algorithm;

[0009] The improved quantum behavior particle swarm optimization algorithm and the cuckoo search algorithm are fused to obtain an improved hybrid cuckoo search quantum behavior particle swarm optimization algorithm;

[0010] Based on the improved hybrid cuckoo search quantum behavior particle swarm optimization algorithm, the variables in the planning layer and the operation layer are solved to complete the configuration.

[0011] Optionally, the process of constructing the ADN source-load-storage model includes:

[0012] The power supply side is modeled based on the output characteristics of wind power generation, photovoltaic power generation and gas turbine power generation; the load side is modeled based on the rigid load response output characteristics and the flexible load response output characteristics; the energy storage model is modeled based on the state of charge of the battery constraints; and the ADN source-load-storage model is constructed based on the constructed power supply side, load side and energy storage side.

[0013] Optionally, the demand response adopts a time-of-use electricity price mechanism in price-driven demand response, and combines interruptible load and transferable load in incentive-driven demand response.

[0014] Optionally, the planning layer is used to obtain the total planning layer cost by summing the investment and construction costs of each device and the results returned by the operation layer planning; then, using power balance constraints, new generator capacity constraints and new energy access ratio constraints as constraints, and minimizing the annual comprehensive planning operation cost as the objective function, the optimal planning scheme for site selection and capacity determination of each device is obtained.

[0015] Optionally, the operation layer is used to obtain the total cost of the operation layer by weighted summing the economic cost of the operation layer, the voltage offset of the distribution network node, the wind power abandonment rate, the photovoltaic abandonment rate and the standard deviation of the equivalent load curve; then, the unit output constraint, the interruptible load constraint, the transferable load constraint, the energy storage device output constraint, the upper power purchase constraint, the power flow constraint and the node voltage constraint are used as constraint conditions, and the total cost of the operation layer is used as the objective function to obtain the optimal scheduling result of the optimal planning scheme of the response planning layer, and the optimal scheduling result includes the daily scheduling curve and the current distribution network operation status;

[0016] Among them, the economic costs at the operating level include the operation and maintenance costs of power supply equipment, the management costs of load demand response, the depreciation costs of energy storage equipment, the annual electricity purchase costs from the superior power grid, and network loss costs.

[0017] Optionally, the process of solving the variables in the planning layer and the operation layer based on the improved hybrid cuckoo search quantum behavior particle swarm optimization algorithm includes:

[0018] Input the parameters of the ADN source-charge-storage model into the improved hybrid cuckoo search quantum behavior particle swarm optimization algorithm to initialize the algorithm parameters;

[0019] Initialize the location and capacity of the power source and energy storage equipment in the ADN source-load-storage model as the initial decision variables of the planning layer;

[0020] Inputting the initial decision variables of the planning layer into the operation layer for solving, and obtaining the optimal solution of the operation layer as the decision variables of the operation layer;

[0021] The decision variables of the operation layer are input into the planning layer for solving, and the initial decision variables of the planning layer are updated to obtain the optimal solution of the planning layer;

[0022] The optimal solution of the operation layer and the optimal solution of the planning layer are input into the planning layer, and the objective function is calculated as the fitness;

[0023] The new fitness is compared with the old fitness, and the Metropolis criterion in the SA algorithm is used to determine whether to accept the new solution, so as to update and iterate the optimal solutions of the planning layer and the operation layer until the maximum number of iterations is reached, and then the optimal solutions of the planning layer and the operation layer are output.

[0024] Optionally, the optimal solution of the planning layer includes:

[0025] The optimal solutions for the number of wind turbine units per unit capacity to be built, the number of photovoltaic power generation units per unit capacity to be built, and the number of gas turbine units per unit capacity to be built constitute a solution set as the optimal planning result.

[0026] Optionally, the optimal solution for the operation layer includes:

[0027] The optimal solutions for the active output of each wind power station, the active output of each photovoltaic power station, and the active output of each gas turbine power station constitute a solution set as the optimal scheduling result.

[0028] The present invention also provides an electronic device, comprising: a memory and a processor; the memory is used to store a program; the processor is used to execute the program to implement each step of the ADN source-load-storage collaborative double-layer optimization configuration method considering demand response.

[0029] The present invention also provides a readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the ADN source-load-storage collaborative double-layer optimization configuration method considering demand response are implemented.

[0030] Compared with the prior art, the present invention has the following advantages and technical effects:

[0031] The present invention establishes a two-layer planning model and realizes the coordination of "planning" and "operation". This method can reduce the investment and operation costs of the active distribution network while improving the reliability, safety and sustainability of the distribution network during operation.

[0032] The improved hybrid cuckoo search quantum behavior particle swarm optimization algorithm proposed in the present invention not only has a strong global optimization capability in the process of solving the bi-level programming problem, thereby avoiding the solution falling into the local optimal solution; it also has higher computational efficiency and saves running time. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0034] Figure 1 A topological diagram of a power distribution system according to an embodiment of the present invention;

[0035] Figure 2 A wind speed and wind power output curve diagram of an embodiment of the present invention;

[0036] Figure 3 1 is a graph showing the time series characteristic of light intensity and temperature in a typical day of an embodiment of the present invention, wherein (a) is a graph showing light intensity, and (b) is a graph showing temperature intensity;

[0037] Figure 4 A photovoltaic power generation output curve diagram of a typical day in an embodiment of the present invention;

[0038] Figure 51 is a graph showing the output characteristic curves of wind power and photovoltaic equipment when DG output is strong or weak according to an embodiment of the present invention, wherein (a) is a graph showing the output characteristic curves of wind power and photovoltaic equipment when DG output is strong, and (b) is a graph showing the output characteristic curves of wind power and photovoltaic equipment when DG output is weak;

[0039] Figure 6 1 is a load power characteristic curve diagram under different scenarios of an embodiment of the present invention, wherein (a) is a load power characteristic curve diagram on weekends, and (b) is a load power characteristic curve diagram on weekdays;

[0040] Figure 7 A two-layer planning framework diagram of an embodiment of the present invention;

[0041] Figure 8 A flowchart of solving a two-level programming model according to an embodiment of the present invention;

[0042] Fig. 9 A load power characteristic curve diagram of demand response according to an embodiment of the present invention;

[0043] Fig.10 It is a graph showing the charging and discharging power and state-of-charge power characteristics of the energy storage device according to an embodiment of the present invention;

[0044] Fig.11 A relationship diagram between the total operating cost and the proportion of interruptible loads according to an embodiment of the present invention;

[0045] Fig.12 It is a convergence curve diagram corresponding to each algorithm of the embodiment of the present invention. DETAILED DESCRIPTION

[0046] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0047] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0048] Embodiment 1

[0049] In this embodiment, a method for optimizing the coordinated two-layer ADN source-load-storage configuration considering demand response is provided. First, the "source-load-storage" system is modeled and analyzed, and the load response analysis of the time-of-use electricity price mechanism in the price-driven demand response is combined. The upper layer of the model is the planning layer, and the lower layer is the operation layer. The two layers have corresponding objective functions and constraints. The interaction between the two-layer models considers the demand response of the ADN source-load-storage coordinated two-layer optimization configuration method, and an improved hybrid cuckoo search quantum behavior particle swarm optimization algorithm is used to solve the two-layer model.

[0050] As a specific embodiment, this embodiment uses the IEEE-33 node power distribution system for case analysis, and its topology is as follows: Figure 1 In this system, the total active load can be 3.715MW, the rated base capacity is 10MVA, the rated voltage is 12.66kV, the node voltage is generally limited to between 0.95pu and 1.05pu, and the branch current is limited to 400A. Other parameters of the system on the power supply side, load side, and energy storage side are shown in Tables 1 to 3.

[0051] Table 1

[0052]

[0053] Table 2

[0054]

[0055]

[0056] Table 3

[0057]

[0058] In the modeling and analysis of the "source-load-storage" system, the power supply side considers the output characteristics of wind power generation, photovoltaic power generation and gas turbine power generation; the load side distinguishes between rigid loads and flexible loads, and determines the transferable loads, shiftable loads and reducible loads in the flexible loads to regulate the load in the optimal way; the energy storage side must consider the charge state of the battery constraints to increase the service life of the energy storage equipment.

[0059] Power supply side modeling and analysis process:

[0060] The inherent variability and intermittent characteristics of wind energy make wind power output show significant randomness and volatility, which undoubtedly increases the uncertainty and difficulty of wind turbine power output. In order to more accurately simulate wind power output and predict its changing trend, this embodiment introduces the Weibull probability density function, which can accurately describe the statistical characteristics of wind speed and comprehensively evaluate and predict the wind speed changes in the target area during the wind farm planning stage, as shown in the following formula:

[0061]

[0062] In the formula, v represents the local wind speed; k and c are the shape parameter and scale parameter, respectively. The shape parameter describes the shape of the distribution and reflects the degree of variation of the wind speed. The scale parameter determines the degree of spread of the distribution, that is, the average level of wind speed. The two can be calculated by the following formulas (2) and (3), respectively:

[0063]

[0064] In the formula, σ w , μ w All can be obtained through historical data.

[0065] Wind power output can be calculated by wind speed and characteristic parameters of wind turbine generator set. According to the internationally accepted power characteristic curve of wind turbine generator set, this embodiment can convert wind speed into power P W The output is shown in formula (4):

[0066]

[0067] Where P Wn is the rated active output of the fan; V i is the cut-in wind speed of the fan; V o is the cut-out wind speed of the fan; V n is the rated wind speed. Specifically, when the wind speed is lower than the cut-in wind speed or higher than the cut-out wind speed, the wind power output is zero. When the wind speed is between the cut-in wind speed and the rated wind speed, the wind power output increases linearly. When the wind speed reaches the rated wind speed but does not reach the cut-out wind speed, the wind power output remains at the rated power.

[0068] Figure 2 The figure shows that the wind speed curve of a typical day is obtained by clustering the wind speed conditions in a certain North China region using the k-means time series clustering algorithm, and the data in the curve is substituted into formula (4) to obtain the wind power output curve of a single wind turbine.

[0069] The output of photovoltaic systems is affected by many factors, including solar radiation intensity, battery module temperature, weather and cloud cover. Changes in these factors can cause large fluctuations in photovoltaic power generation and have a certain degree of randomness. Among them, sunlight is one of the key factors affecting photovoltaic power generation. Its variability and uncertainty bring challenges to the prediction of photovoltaic power generation output. Therefore, the output research of photovoltaic power stations requires the analysis of solar radiation intensity. Figure 3 shown.

[0070] This embodiment is based on the light intensity distribution obtained by the Beta distribution model. The Beta distribution is a distribution suitable for describing variables between 0 and 1, and can fit different types of distributions well. In this embodiment, the sunlight intensity is normalized to between 0 and 1, and the Beta distribution is used to model its probability distribution as shown in formula (5):

[0071]

[0072] In the formula, r PH is the solar intensity at a certain time section; r PHmax is the maximum historical solar intensity of the time section; f(rPH) represents r PH The corresponding probability density function; α PH and β PH is the shape parameter of the Beta distribution. By choosing a suitable α PH and β PH Parameters, Beta distribution can well simulate the probability distribution of sunshine intensity in different time periods, and then be used to calculate the expected photovoltaic power output in different time periods, as shown in equations (6) and (7):

[0073]

[0074] In the formula, μ PH and σ PH Respectively represent the mean and standard deviation of historical data of solar intensity.

[0075] The output of photovoltaic power generation can be characterized as a function with temperature and radiation intensity as independent variables. The corresponding output characteristic function is shown in formula (8):

[0076]

[0077] Where P pv Indicates that the photovoltaic power station is under the condition of sunlight intensity G c Active power output at time G r and P r is the radiation intensity and maximum output power under standard test conditions (the temperature is 25°C), k is the power temperature coefficient, Tc and T r are the photovoltaic cell temperature and the reference temperature respectively.

[0078] Figure 2 Shown are the time series characteristics of light intensity and temperature in a typical day.

[0079] Substituting the temperature curve and light intensity curve in a typical day into equation (8), we can obtain the photovoltaic output characteristics in a typical day as follows: Figure 4 shown.

[0080] According to the strength of distributed generation (DG) output, DG strong output and DG weak output are generated respectively. Under this characteristic, two sets of wind power output and photovoltaic output characteristic per unit value curves are generated, with the rated capacity of the equipment as the reference value, such as Figure 5 (a) and Figure 5 (b) as shown.

[0081] The historical load data of a certain area in North China were selected and clustered to obtain the load output characteristic curves under the two scenarios of working days and rest days. The annual maximum power of the load in each scenario was taken as the benchmark value. Figure 6 (a) and Figure 6 (b) as shown.

[0082] It is assumed that the DG high output and DG low output time account for 1 / 2 of the total time, weekdays and weekends account for 5 / 7 and 2 / 7 respectively, and it is assumed that the DG output is irrelevant to whether it is a weekday. In this way, four scenarios are divided according to the DG output and the user's work and rest time. The scenario settings and their corresponding probabilities are shown in Table 4:

[0083] Table 4

[0084]

[0085] Combined with local climate conditions and load characteristics, for the equipment to be planned, photovoltaic equipment is preferentially built near nodes with good lighting conditions, so {7,13,18,28,32} are determined as potential installation nodes. Considering the characteristics of wind resources, nodes close to the edge of the system with small loads but abundant wind resources are selected as candidate installation nodes, thus {6,15,19,28,32} are selected as candidate installation locations for wind power generation equipment. As a peak-shaving resource, gas turbines can be installed at load centers or nodes with large system capacity, and {8,10,23,24,30} are selected as candidate nodes. At the same time, considering the key role of energy storage equipment in improving system stability and reliability, it can be installed near nodes with large load or new energy generation fluctuations, and {4,11,15,21,23,31} are selected as candidate nodes. In addition, {3,9,16,22,25,29} can be selected as installation nodes for flexible loads. The unit capacity of photovoltaic generator sets, wind generator sets and gas turbine generator sets is uniformly set to 20kW, and the upper limit of installation of each node is 1000kW. The 24-hour time-of-use electricity price data of the selected areas are shown in Table 5:

[0086] Table 5

[0087]

[0088] Micro gas turbines can generate high-temperature and high-pressure gas through natural gas combustion to drive the turbine to rotate and generate electricity, thereby realizing the conversion of chemical energy-thermal energy-mechanical energy-electrical energy. The gas turbine output model is shown in formula (9):

[0089] P gas (t) = V gas (t)η gas θ gas (9)

[0090] Where P gas (t) is the power of the gas turbine generator set; V gas (t) is the natural gas consumption at time t, η gas is the power generation efficiency of the gas turbine, θ gas is the power angle size.

[0091] In this embodiment, the micro gas turbine is used as a type of fully dispatchable distributed power source, and its output range is shown in formula (10):

[0092] 0≤P gas ≤S gasmax (10)

[0093] Where P gas is the active power output of the gas turbine; S gasmax is the installed capacity of the gas turbine.

[0094] Load side modeling and analysis process:

[0095] In terms of participating in grid regulation, user loads can be divided into two categories: rigid loads and flexible loads. Compared with rigid loads, flexible loads can be actively adjusted and responded through dynamic electricity price adjustments and policy-driven methods, and adjust power consumption characteristics within a specified time, so as to better adapt to the operation needs of the grid.

[0096] Flexible loads can be divided into transferable loads, shiftable loads and curtailable loads according to the characteristics of their interaction with the power grid. Transferable loads can flexibly adjust the characteristics and time of electricity consumption while keeping the total electricity consumption unchanged. Shiftable loads are limited by the production process and can only shift electricity loads in different time periods. Interruptible loads can reduce electricity consumption within a specific time.

[0097] This embodiment does not involve industrial production issues, so only the application of transferable loads and interruptible loads is considered. According to the output plan of each user, the total response output of the flexible load per unit time can be obtained:

[0098]

[0099] In the formula, S n,t represents the response state of user n at time t. If it is 1, it means the load increases, -1 means the load decreases, and 0 means the load remains unchanged. N represents the number of users that can participate in the scheduling. P n,t represents the power of user n at time t.

[0100] Energy storage side modeling and analysis process:

[0101] This embodiment uses a battery to represent the energy storage system. During the charging and discharging process of the battery energy storage system, the charge state of the energy storage system in adjacent time periods is shown in formula (12):

[0102]

[0103] In the formula, SOC E，t is the state of charge of the initial energy storage at time t; E ESS is the rated capacity of the energy storage device; D ESS is the self-loss coefficient of the energy storage system; P char,t and P dis,t are the charging and discharging power of the energy storage device in period t respectively; η char and η dis are the charging and discharging efficiency of the energy storage device; U Ech,t and U Edis,t Respectively represent the charging and discharging states of the energy storage device, which are 0 and 1 variables. The charging and discharging states of the energy storage system cannot exist at the same time, and the following constraints need to be met:

[0104] U Ech,t +U Edis,t ≤1 (13)

[0105] In addition, in order to improve the life of energy storage equipment, the charging and discharging state of the battery needs to meet the following constraints:

[0106] SOC Emin ≤SOC E，t ≤SOC Emax (14)

[0107] P Emin ≤P char,t ≤P Emax (15)

[0108] P Emin ≤P dis,t ≤P Emax (16)

[0109] In the formula, SOC Emax With SOC Emin Respectively represent the maximum and minimum values ​​of the energy storage system’s state of charge; P Emin With P Emax They represent the minimum and maximum charging and discharging power of the energy storage system respectively.

[0110] This embodiment is based on the characteristics of the active distribution network, applies the time-of-use electricity price mechanism in the price-driven demand response, and considers resources such as interruptible loads and transferable loads in the incentive-driven demand response, so as to make full use of the initiative of flexible loads.

[0111] Furthermore, Price-Driven Demand Response (PDR) mainly uses the difference in electricity prices in different time periods to encourage users to adjust their electricity consumption behavior according to electricity price signals. This model covers mechanisms such as time-of-use electricity prices and real-time electricity prices. In time-of-use electricity prices, the electricity price level is set according to the load of the power grid, aiming to encourage users to increase electricity consumption during low electricity price periods and reduce usage during high electricity price periods. Real-time electricity prices are more flexible and adjust electricity prices according to the real-time supply and demand conditions of the market, thereby providing users with more accurate electricity price period information to achieve more detailed load adjustments.

[0112] Incentive-Driven Demand Response (IDR) mainly encourages users to actively reduce electricity consumption when the grid load is high through economic incentives. Due to its design and implementation mechanism, IDR can usually provide a faster load regulation response.

[0113] Figure 7What is shown is the two-level planning framework diagram of the source-load-storage coordinated two-level optimization configuration model, which reflects the relationship between the decision-making models and decision variables of the upper planning layer and the lower operation layer under different situations.

[0114] The upper layer is the planning layer, which aims to obtain the optimal planning scheme by planning the site selection and capacity of the equipment, with the goal of minimizing the annual comprehensive planning and operation cost. The objective function is mainly divided into the investment and construction cost of each equipment and the results returned by the operation layer planning. The upper layer constraints include power balance constraints, new generator capacity constraints, and new energy access ratio constraints;

[0115] Specifically, the objective function of the upper planning layer is to minimize the total cost, as shown in the following formula:

[0116] C total =C ght +f deco (17)

[0117] Where C ght is the investment and construction cost of each equipment, including distributed power generation stations, gas turbine units and energy storage equipment, etc. deco The results returned by the lower layer (operation layer) planning.

[0118]

[0119] In the formula, are the investment and construction costs of wind power stations, photovoltaic power stations, gas turbine generator sets, and energy storage equipment, respectively. The corresponding calculation formulas are shown in equations (19) to (22):

[0120]

[0121] In the above formula, a W ,a PH ,a gas ,a E are the investment discount rates of the wind power station, photovoltaic power station, gas turbine power station and energy storage equipment to be built; d W ,d PH ,d gas ,d E N are the service life of wind power station, photovoltaic power station, gas turbine power station and energy storage equipment respectively; W ,N PH ,N gas ,N E are the number of candidate locations for newly built wind power stations, photovoltaic power stations, gas turbine generator sets, and energy storage equipment; C IW , C IPH , C IgasE is the unit capacity cost of the i-th newly built wind power station, photovoltaic power station, and gas turbine generator set respectively; W0 、E PH0 、E gas0 are the new unit capacities of wind power stations, photovoltaic power stations, and new gas turbine generator sets; C IEP,i , C IES,i are the unit power cost and unit capacity cost of the i-th energy storage device respectively; P E0 , S E0 are the unit energy storage power and unit capacity of the i-th energy storage device to be built; v W,i ,v PH,i ,v gas,i are the unit capacities of the newly built wind power station, photovoltaic power station, and gas turbine generator set in the i-th place; v PE,i ,v SE,i are the unit energy storage power and unit energy storage capacity of the i-th energy storage device respectively.

[0122] The upper planning layer constraints include power balance constraints, new generator capacity constraints, and new energy access ratio constraints. The power balance constraint is shown in the following formula:

[0123]

[0124] Where P W,i , P PH,i , P gas,i are the output of the i-th wind power generation, photovoltaic power generation, and gas turbine power station respectively; P E,i is the charging or discharging power of the i-th energy storage device, when P E,i When the value is positive, the energy storage device discharges outward; when P E,i When it takes a negative value, the energy storage device is in a charging state; P load,i is the total load on the electricity consumption side of the i-th node; N node is the total number of active distribution network nodes.

[0125] The capacity constraints of the newly built generator sets are shown in the following formula, which are the capacity constraints of the newly built gas turbine generator sets, the capacity constraints of the newly built wind power stations and the capacity constraints of the newly built photovoltaic power stations:

[0126]

[0127] In the formula, E gas0 、E W0 、E PH0 They are the new unit capacities of new gas turbine power stations, wind power stations, and photovoltaic power stations; They are the maximum new capacities of gas turbine power stations, wind power stations and photovoltaic power stations respectively.

[0128] Constraints on the proportion of new energy access:

[0129]

[0130] In this embodiment, the upper and lower limits of the proportion of renewable energy power generation are set at 20% to 50% respectively, taking into account the requirements of the dual-carbon policy, the availability of new energy resources, the technical capabilities of the distribution network, and the overall demand, stability and flexibility of the power system.

[0131] The lower layer is the operation layer, which aims to optimize the scheduling and demand response management of the active resources on the source-load-storage side of ADN based on the planning results of the upper layer, and obtain the daily scheduling curve to respond to the planning requirements of the upper layer and the current operation of the distribution network. The objective function mainly considers the economic cost and power quality indicators of the distribution network operation, coordinates the output of various power supply equipment, flexible loads, and energy storage equipment and the amount of electricity purchased from the superior power grid, so as to achieve the purpose of reducing operating costs, reducing network losses, shaving peaks and filling valleys for loads, and improving voltage stability and distributed power supply absorption capacity. The objective function includes the operation and maintenance of power supply equipment Cost, management cost of load demand response, depreciation cost of energy storage equipment, annual electricity purchase cost from the superior power grid, network loss cost, distribution network node voltage offset, wind power abandonment rate and photovoltaic abandonment rate, standard deviation of equivalent load curve; the lower-level constraints include unit output constraints, interruptible load constraints, transferable load constraints, energy storage equipment output constraints, superior power purchase constraints, power flow constraints and node voltage constraints; the upper planning layer provides the basis for equipment configuration and parameter setting for the lower operation layer, while the lower operation layer feeds back the actual operation results to the upper layer for the upper layer to make planning adjustments.

[0132] Furthermore, its objective function includes the following parts:

[0133] Operation and maintenance costs of power supply equipment:

[0134] During the operation of ADN, the operation and maintenance cost of power supply equipment is shown in formula (28):

[0135]

[0136] In the formula, C YWP is the total operation and maintenance cost of all power equipment; and are the operation and maintenance costs of wind power generation equipment, photovoltaic power generation equipment, and gas turbine power generation equipment respectively; S is the number of days of typical day scenarios; is the probability of occurrence of each typical day; C OP,W , C OP,PH , C OP,gas are the operation and maintenance costs of wind power generation equipment, photovoltaic power generation equipment, and gas turbine power generation equipment corresponding to unit power generation; P W,s,i,t , PPH,s,i,t , P gas,s,i,t They are respectively the active output of each wind power station, photovoltaic power station and gas turbine power station of the ith wind power generation equipment, photovoltaic power generation equipment and gas turbine power generation equipment at time t under the typical day scenario s; Δt is the duration, which is 1h in this model.

[0137] Management costs of load demand response:

[0138] In this embodiment, while using interruptible loads and transferable loads to maintain the stable operation of the distribution network, it is necessary to provide certain subsidies to the users corresponding to the loads, as shown in formula (32):

[0139]

[0140]

[0141] In the formula, C loadT The total load demand response management cost; and are load transfer compensation cost and load dispatch compensation cost respectively; C OLT,i and C OLI,i are the unit load transfer compensation cost and the unit load interruption compensation cost respectively; P LT,s,i,t and P LI,s,i,t They are respectively the transfer load and interruption load of the demand response node at the i-th node at time t under the typical day scenario s.

[0142] Depreciation cost of energy storage equipment:

[0143] The depreciation cost of energy storage equipment is shown in formula (35):

[0144]

[0145] Where P E,s,i,t represents the charging and discharging power of the i-th energy storage device at time t in scenario s; U Edis,s,i,t and U Ech,s,i,t are state variables, representing the discharge and charge states of the i-th energy storage device at time t under scenario s, and both are 0 and 1 variables.

[0146] Annual electricity purchase cost from the upper grid:

[0147]

[0148] In the formula, C buy is the cost of purchasing electricity from the active distribution network to the upper grid; C price,s,t is the electricity price at time t in scenario s; P buy,s,tIt is the power obtained by the active distribution network from the upper distribution network at time t.

[0149] Network loss costs:

[0150]

[0151] In the formula, is the network loss of the active distribution network during operation; P line,s,i,t is the active power of the line flow, C nloss is the network loss cost per unit power of the transmission network.

[0152] The costs in (28)-(37) above can be summarized as the operating layer economic cost f deco , as shown in formula (38):

[0153]

[0154] Distribution network node voltage offset:

[0155] This embodiment takes the minimum total offset of the node voltage as part of the objective function:

[0156]

[0157] In the formula, f dU is the total offset of the voltage of each node in the active distribution network; is the node voltage per unit value of the i-th node at time t in each scenario.

[0158] Wind power abandonment rate and photovoltaic abandonment rate:

[0159] In order to ensure the consumption of distributed power sources, this embodiment sets the wind power abandonment rate f LW and the PV abandonment rate f LPH Incorporate into the overall objective function, as shown in equations (40) and (41):

[0160]

[0161] Where, T W and T PH are the theoretical available hours of wind power and photovoltaic power, respectively; P WF,s,i,t and P W,s,t,i are the wind power forecast value and actual output power value of the ith node at time t under scenario s; P PHF,s,t,i and P PH,s,t,i They are respectively the predicted photovoltaic power value and actual output power value of the i-th node at time t under scenario s.

[0162] Standard deviation of the load curve:

[0163] In order to measure the effect of demand-side response on reducing load fluctuations and achieving peak shaving and valley filling, this embodiment calculates the standard deviation of the net load curve, as shown in formula (42):

[0164]

[0165] In the formula, f sd is the standard deviation of the equivalent load curve, P JLoad,s,i,t represents the net load of the ith node at time t in scenario s, P Load,s,i,t It represents the total power load of the i-th node at time t in scenario s.

[0166] In summary, the calculation formula of the total objective function of the operation layer is shown in formula (43):

[0167] f down =ω1f deco +ω2f dU +ω3f LW +ω4f LPH +ω5f SD (43)

[0168] In the formula, ω1, ω2, ω3, ω4, and ω5 are the weight coefficients of each sub-objective function. The weight coefficients in the lower-level objective function of this example are shown in Table 6, where ω1 is dimensionless, and ω2, ω3, ω4, and ω5 are all in ten thousand yuan:

[0169] Table 6

[0170]

[0171] The lower level constraints also consist of the following parts:

[0172] Unit output constraints:

[0173] The upper and lower limits of wind turbine output are shown in formula (44):

[0174] 0≤P W,s,i,t ≤E W0 v W,i i=1,2,...N W (44)

[0175] The upper and lower limits of the photovoltaic generator set output are shown in formula (45):

[0176] 0≤P PH,s,i,t ≤E PH0 v PH,i i=1,2,…N PH (45)

[0177] The upper and lower limits of the gas turbine generator set output are shown in formula (46):

[0178] 0.5E gas0 v gas,i ≤P gas,s,i,t ≤E gas0 v gas,i i=1,2,…N gas (46)

[0179] Constraints for interruptible loads:

[0180] 0≤P LI,s,i,t ≤P LI,s,i,MAX (47)

[0181]

[0182] Where P LI,s,i,MAX is the maximum active power of the interruptible load at node i under scenario s; It is the power percentage of interruptible load in the total load.

[0183] Constraints on transferable loads:

[0184] Transferable load power constraints:

[0185] P LT,s,i,MIN ≤P LT,s,i,t ≤P LT,s,i,MAX (49)

[0186] Where P LT,s,i,MAX and P LT,s,i,MIN They represent the maximum power and minimum power limits of the transferable load at node i at time t under scenario s, respectively;

[0187] In addition, the transferable load satisfies the load transfer conservation constraint within a scheduling cycle:

[0188]

[0189] Energy storage equipment output constraints:

[0190] The energy storage device charging / discharging power constraint and device SOC constraint are shown in equations (51) and (52), respectively;

[0191] P E,s,j,t ≤v PE,i P E0 (51)

[0192] SOC min ≤SOC E,t ≤SOC max (52)

[0193] Constraints on power purchase from higher authorities:

[0194] P buy,s,t ≥0 (53)

[0195] Power flow constraints:

[0196]

[0197] Where P i t and Respectively represent the active and reactive power values ​​of node i; U i,t and U j,t are the voltage amplitudes of nodes i and j respectively; G ij and B ij is the admittance between node i and node j; θ ij is the phase angle difference between node i and node j.

[0198] Node voltage constraints:

[0199]

[0200] In the formula, and Represent the voltage amplitude of nodes i and j respectively; r ij and x ij is the impedance between node i and node j.

[0201] In the two-layer model constructed in this embodiment, the planning layer and the operation layer do not exist in isolation, but achieve global optimization by interacting with each other and transmitting decision information to each other.

[0202] The upper layer, as the planning layer, optimizes the construction plan of distributed power generation, the construction plan of gas turbine power station and the site selection and capacity determination plan of energy storage, and obtains the optimal solution of variables such as the number of wind turbine units per unit capacity to be built, the number of photovoltaic power generation units per unit capacity to be built, and the number of gas turbine units per unit capacity to be built. The solution set X is passed to the lower operation layer as the optimal planning result, and affects the objective function and constraints of the lower problem.

[0203] The lower layer is the operation layer. Based on the upper-level planning scheme, it takes the operation economic cost, wind and solar power consumption, voltage and load stability of each operation scenario as the objective function. It coordinates and optimizes the output distribution of distributed power sources and conventional thermal power units, demand response control of flexible loads, coordinated control of multiple energy storage devices, and power purchase behavior based on time-of-use electricity prices. It obtains the optimal scheduling results of various active resources, the optimal solutions of variables such as the active output of each wind power station, the active output of each photovoltaic power station, and the active output of each gas turbine power station, and the solution set Y is fed back to the upper-level planning model to update the corresponding objective function value.

[0204] This embodiment uses an improved hybrid cuckoo search quantum behavior particle swarm optimization algorithm (ICSQPSO algorithm):

[0205] The particle swarm optimization algorithm (PSO) finds the optimal solution to the problem by simulating the process of a flock of birds flying and searching in the solution space. In the PSO algorithm, each "particle" represents a potential solution in the solution space and is assigned a speed to determine the direction and step size of its search. The speed update and position update formulas of the particle are shown in equations (57) and (58), respectively:

[0206]

[0207] Quantum-behaved Particle Swarm Optimization (QPSO) allows particles to appear anywhere in the search space with a certain probability, thereby increasing the chance of finding the global optimal solution.

[0208] As an additional embodiment, the particle position update formula corresponding to QPSO is shown in (59):

[0209] x i (t+1)=P i (t)±β·|M(t)-x i (t)|·ln(1 / u) (59)

[0210]

[0211] Where P i (t) is the best position found by particle i in history, i.e., the individual optimal solution; β is the contraction-expansion coefficient, which is used to control the size of the particle search step, usually set as a constant, such as β = 0.5 or dynamically adjusted through some strategy; M(t) is the average position of the optimal solution of all individual particles, called the intention center of the particle group; M(t)-x i (t) represents the distance between the current position of particle i and the intended center; ln(1 / u) is a random number factor used to generate a step length distribution that conforms to the Lévy flight, where u is a uniform random number in the interval [0,1], and “±” represents that the particle can search in any direction on both sides of the individual optimal position; N L is the size of the particle group.

[0212] The Cuckoo Search algorithm (CS) is an optimization algorithm inspired by the parasitic reproduction behavior and Lévy flight behavior of cuckoos in nature. The position update of the Cuckoo Search algorithm is mainly achieved through Lévy flight, as shown in formula (61):

[0213]

[0214] In the formula, is the position of the i-th cuckoo at the t-th iteration; α bc is the step size scaling factor, which is used to control the size of the step size. represents the Hadamard product, and Levy(λ) is the Lévy distribution used to generate the step size of the Lévy flight.

[0215] The main steps of the standard CS algorithm are as follows:

[0216] (1) Algorithm initialization: Randomly generate the initial cuckoo (solution) position.

[0217] (2) Lévy flight update: According to the above formula, the cuckoo's position is updated through the Lévy flight formula, that is, a new solution is generated.

[0218] (3) Determination of the new solution: Upgrade to generate a random number r find ∈[0,1] and the probability of egg abandonment P find For comparison, if r find Greater than P find , the newly generated solution is updated to the optimal solution, otherwise the solution is eliminated.

[0219] (4) The cuckoo will re-perform the Lévy flight to update its position.

[0220] (5) Iterate until convergence: Repeat steps 2 to 4 until the maximum number of iterations is met or the convergence condition is reached.

[0221] Simulated Annealing (SA) is a probabilistic global optimization algorithm that simulates the annealing process of matter to find the global optimal solution of a given function in a huge search space.

[0222] The specific process of the SA algorithm is as follows:

[0223] (1) Initialization parameters: Select a higher initial temperature T high , initialize the solution state X0, and set the termination temperature T end .

[0224] (2) Add perturbation, that is, starting from the current solution X, a new candidate solution X′ is generated through some perturbation mechanism.

[0225] (3) Calculate the energy difference, that is, calculate the energy difference ΔE=E(X′)-E(X) between the new solution X′ and the current solution X.

[0226] (4) Judge according to the Metropolis criterion, that is: if ΔE>0, the new solution is worse than the current solution, and the new solution X′ is directly accepted; if ΔE<0, it can be judged that the new solution is not as good as the current solution or is equal to it. At this time, there is still a certain probability to accept the new solution, and this probability P ME =e -ΔE / T(t) , where T(t) is the temperature at time t. This step allows the algorithm to accept poor solutions to a certain extent, thus making it possible to escape the trap of local optimal solutions.

[0227] (5) Cool down, and the temperature is updated as shown in formula (62):

[0228] T(t+1)=χT(t) (62)

[0229] Where χ is a cooling coefficient less than 1, which is used to gradually reduce the temperature.

[0230] (6) Termination condition: When the temperature drops to T end The algorithm stops when the optimal solution is not found after multiple consecutive iterations or other termination conditions are met.

[0231] The ICSQPSO algorithm used in this embodiment is based on the QPSO algorithm and incorporates the Metropolis standard in the SA algorithm, allowing for the acceptance of poor solutions with a certain probability during the iterative exploration process. This strategy effectively avoids the algorithm from falling into the dilemma of local optimality in the early stages, significantly enhances the ability to find the best solution globally, and innovatively optimizes the step size mechanism of the CS algorithm. By introducing a dynamic adjustment strategy based on the Lévy distribution to improve the Lévy flight, the algorithm can flexibly switch between macroscopic search and microscopic fine search when exploring the solution space.

[0232] As an implementable method, the process of solving the double-layer model using the improved hybrid cuckoo search quantum behavior particle swarm optimization algorithm adopted in this embodiment includes:

[0233] The improved hybrid cuckoo search quantum-behaved particle swarm optimization algorithm (ICSQPSO algorithm) first upgrades the traditional particle swarm optimization algorithm (PSO) to a quantum-behaved particle swarm optimization algorithm (QPSO) by introducing the concepts of quantum states and probability clouds, giving particles the characteristic of existing in the solution space in a probabilistic form; secondly, the algorithm incorporates the Metropolis criterion in the SA algorithm, allowing poor solutions to be accepted with a certain probability during the iterative exploration process, avoiding the algorithm from falling into the trap of local optimality in the early stages; finally, the ICSQPSO algorithm also innovatively optimizes the step size mechanism of the Cuckoo Search algorithm (CS) algorithm, and improves the Lévy flight by introducing a dynamic adjustment strategy based on the Lévy distribution, so that the algorithm can flexibly switch between macroscopic search and microscopic fine search when exploring the solution space.

[0234] As an implementable method, the specific process of using the ICSQPSO algorithm to solve the model of this embodiment is as follows: Figure 8 The operation steps are as follows:

[0235] Step 1: Input the active distribution network system parameters and the typical daily curves of DG and load under each scenario;

[0236] Step 2: Assign values ​​to the parameters such as the maximum number of iterations, initial temperature, end temperature, temperature attenuation coefficient, etc. in the standard SA algorithm, and initialize the iteration coefficient t1 = 0;

[0237] Step 3: Initialize the upper model and parameters. Initialize the location and capacity of the power source and energy storage device in the upper model (including v W,i ,v PH,i ,v gas,i , v PE,i v SE,i etc.), as the initial decision variable X(t1) to the planning layer, to initialize the number, position and speed of particles in the CSQPSO algorithm corresponding to the upper model, and assign values ​​to parameters such as the contraction-expansion coefficient, compensation scaling factor, and maximum number of iterations;

[0238] Step 4: Initialization of CSQPSO algorithm parameters in the lower model. Define the size of the particle swarm and set the particle positions, including the output of each generator set, flexible load, energy storage equipment, and purchased power, i.e., P W,s,i,t , P PH,s,i,t , P line,s,i,t , P gas,s,i,t , P LT,s,i,t , PLI,s,i,t , P E,s,j,t , P buy,s,t , And assign values ​​to parameters such as the shrinkage-expansion coefficient, the compensation scaling factor, and the maximum number of iterations;

[0239] Step 5: Substitute the planning layer decision variable X(t1) into the lower-level planning model, call the CSQPSO algorithm in Section 3.3.2 to solve the lower-level model, and obtain the optimal solution Y(t1) of the lower-level model as the decision variable of the operation layer;

[0240] Step 6: Substitute the optimal solution Y(t1) of the lower model into the upper model for solution, and update the solution X(t1) of the upper model by calling the CSQPSO algorithm;

[0241] Step 7: Substitute X(t1) and Y(t1) into the upper-level planning model, calculate the objective function and use it as the fitness;

[0242] Step 8: Compare the newly obtained fitness value with the old fitness value, and use the Metropolis criterion mentioned in Section 3.5.3 to determine whether to accept the new solution. If accepted, update the optimal solution and record the upper and lower layer decision variables at this time; otherwise, keep the original solution;

[0243] Step 9: Call the standard SA algorithm and call equation (62) to complete the annealing process. If t1 ≥ M th And T(t1)≤T f (When the maximum number of iterations is reached and the SA algorithm cools down to the termination temperature), go to step 10, otherwise go to step 5 and set t1 = t1 + 1;

[0244] Step 10: Output the group optimal solution. At this time, the solution of the upper model is used as the optimal planning solution, and the solution of the lower model is used as the optimal scheduling strategy. The calculation is completed.

[0245] The ICSQPSO algorithm parameter settings for solving this example are shown in Table 7:

[0246] Table 7

[0247]

[0248] In order to study the impact of load demand response and the introduction of energy storage equipment on the comprehensive cost of the configuration scheme, this embodiment provides the following comparison scheme:

[0249] (1) Scheme 1 (the scheme proposed in this embodiment): adopt the source-load-storage coordinated planning method established in Chapter 3, optimize the configuration of DG, MT, and ESS at the same time, and consider the load demand response.

[0250] (2) Scheme 2: Compared with Scheme 1, the total load remains unchanged, the load demand response is not considered, and the DG, MT, and ESS are optimized at the same time. The other parameters are the same as those of Scheme 1.

[0251] (3) Scheme 3: Compared with Scheme 1, the optimization configuration of DG and MT is performed simultaneously, the energy storage equipment is not considered, and the load demand response is considered. The other parameters are the same as those of Scheme 1.

[0252] (4) Scheme 4: Compared with Scheme 1, the optimal configuration of DG and MT is performed, energy storage equipment is not considered, and the load demand response is not considered (the total load remains unchanged). The other parameters are the same as those of Scheme 1.

[0253] The simulation calculation and verification were performed by MatlabR2020a programming, and the optimal solution of source-load-storage coordinated planning was obtained. Among them, the power supply equipment and site selection and capacity planning results corresponding to Scheme 1 are shown in Table 8:

[0254] Table 8

[0255]

[0256] Under this scheme, the economic costs and various operating indicators generated during all planning and operation processes are shown in Table 9:

[0257] Table 9

[0258]

[0259]

[0260] From the optimal planning scheme listed in Table 8, we can see that the number of distributed power generation units installed is significantly higher than that of gas turbine power generation equipment. This is mainly due to the fact that the investment and construction costs of distributed power generation units are significantly lower than the cost of distribution system operators purchasing electricity from the upper power grid. Therefore, increasing the penetration rate of photovoltaics in the distribution system will effectively save overall social costs. The power generation and operation and maintenance costs of gas turbines are expensive, so they are only configured in small quantities as emergency power sources during peak load moments.

[0261] In the results obtained, the number of photovoltaic power generation equipment is significantly higher than the number of wind power generation equipment. This is mainly because wind power output shows a certain anti-peak regulation characteristic, especially during the period of low load at night. Excessive investment may lead to waste of wind power resources. On the contrary, photovoltaic output is mainly concentrated in the daytime period with sufficient sunshine. This will not only not aggravate the peak-to-valley difference of equivalent load, but will better promote the consumption and utilization of electricity. In the second plan, it can be seen that photovoltaic power stations are configured on all candidate nodes. Through this decentralized and fully covered configuration method, it is possible to effectively avoid the situation where branch current exceeds the limit during the peak moment of photovoltaic output, and increase the access ratio of photovoltaic capacity.

[0262] In order to analyze the load demand response and the charging and discharging of the energy storage equipment, the DG load at node 9 and the energy storage equipment at node 11 in scenario 1 are taken as examples. At this time, on this typical day, the real-time charging and discharging per unit curve of the energy storage (with the rated capacity of the energy storage equipment as the reference value) and the load time series per unit curve obtained after the load demand response (with the annual maximum power of the load as the reference value) are as follows: Fig. 9 and Fig.10 As shown. Combined Fig. 9 , it can be seen that demand response can play the role of smoothing the load curve; Fig.10 It can be seen that the energy storage equipment discharges during peak electricity consumption and charges during low electricity consumption, which also has the effect of reducing peak power and filling valley power.

[0263] In order to verify the superiority of the proposed method in planning and operation, this embodiment solves the four proposed schemes, and the site selection and capacity determination results of each source-load-storage unit are shown in Table 10:

[0264] Table 10

[0265]

[0266] In Table 10, the numbers outside the brackets indicate the node number, and the numbers in the brackets indicate the number of devices installed per unit capacity at a certain node.

[0267] At this time, the economic costs corresponding to each plan are shown in Table 11:

[0268] Table 11

[0269]

[0270] The active distribution network operation indicators corresponding to each scheme are shown in Table 12:

[0271] Table 12

[0272]

[0273] It can be seen from Tables 11 and 12 that, considering various economic costs and operating indicators, Scheme 1 is the best, followed by Scheme 2 and Scheme 3. Scheme 4 only coordinates the planning and operation of distributed power sources through the use of gas turbines and purchasing electricity from the superior distribution network. At the same time, it requires the most expensive regulation cost, and all operating indicators are not ideal.

[0274] Comparing the planning results in Scheme 1 and Scheme 2, it can be seen that the proportion of distributed power sources in Scheme 1 is significantly higher than that in Scheme 2. This is because the access of flexible loads significantly improves the system's ability to absorb distributed power sources, especially photovoltaic power generation. At this time, the proportion of gas turbine power generation and the amount of electricity purchased from the upper power grid are significantly reduced, and the system's planning and operating costs are reduced.

[0275] Comparing the planning results in Scheme 1 and Scheme 3, it can be seen that the introduction of energy storage equipment can significantly reduce network losses and load fluctuations. At the same time, comparing Scheme 1, Scheme 2 and Scheme 3, it can be seen that because the investment cost of energy storage batteries is relatively high, when the optimization effects of various operating indicators are similar, the cost-effectiveness of using flexible loads for optimization is better than that of using energy storage batteries for optimization.

[0276] By comparing the results of Scheme 1 and Scheme 4, it can be seen that the source-load-storage coordinated planning of multiple types of power sources, energy storage batteries and flexible loads not only reduces the comprehensive construction and operation costs, but also improves the absorption rate of distributed power sources, stabilizes the node voltage, and plays a role in peak shaving and valley filling for the load, thereby improving the operating performance of the active distribution network.

[0277] In order to further explore the impact of demand response on collaborative planning, this embodiment takes the interruptible load in the DR load as an example, and increases the maximum response capacity proportion of the interruptible load from 0 to 50%, while other parameters remain unchanged. Fig.11 Shown is the relationship between the total cost of simulated operation and the demand response capacity.

[0278] Depend on Fig.11It can be seen that the increase in the proportion of DR capacity can reduce the comprehensive operating cost of the system within a certain range. This is because the load can realize the consumption of distributed power sources through the demand response mechanism, reduce network losses, and increase the stability of voltage and load. However, as the maximum proportion of load demand response increases further, the total operating cost remains unchanged, which means that when the proportion of interruptible load capacity exceeds a certain threshold (here is 30%), the excessive regulation of DR increases the total operating cost. This is because DR's regulation ability for active distribution networks is limited, and excessive regulation will cause a certain degree of damage to the interests of users of flexible loads. At this time, operators need to pay more compensation fees. At this time, other means (such as the dispatch of energy storage equipment) can be used to replace the use of load demand response. This result also provides a certain reference for the use of flexible loads in the coordinated planning and operation of source-load-storage in active distribution networks.

[0279] In order to verify the superiority of ISCQPSO algorithm in the optimization process, ICSQPSO algorithm, standard SA algorithm, standard CS algorithm, QPSO algorithm and PSO algorithm were used to solve the proposed model respectively. Fig.12 Shown is a comparison of the convergence curves corresponding to each algorithm.

[0280] Depend on Fig.12 It can be seen that the traditional PSO algorithm has a faster solution speed, but the comprehensive cost of the corresponding solution is higher because it quickly falls into the local optimal solution; the solution obtained by the QPSO algorithm is more ideal, which shows that it may have an advantage over PSO in finding the global optimal solution, but there is still a certain possibility of falling into the local optimal solution; the curve of the SA algorithm shows a rapid initial decline and a slow decline in the later period, reflecting the characteristics of "accepting poor solutions" in its simulated annealing process, which helps to jump out of the local optimal solution, and the comprehensive cost corresponding to the obtained result is also lower; the final planning solution corresponding to the curve of the standard CS algorithm has a more ideal result, which means that the cuckoo search algorithm can continuously find better solutions instead of falling into the local optimal solution at an early stage, but at the same time, the algorithm needs more iterations to find the optimal solution, and it takes a long time to solve complex planning problems. The ICSQPSO algorithm basically does not fall into the situation of having the optimal solution, and the comprehensive total cost of the optimal solution obtained is the lowest, that is, the optimal operation plan is not missed because of staying in the local optimal solution; and the algorithm is faster to solve the model operation speed, and it can converge in only 8 iterations. In summary, the hybrid optimization method proposed in this embodiment has better optimization performance.

[0281] The two-level planning model established in this embodiment realizes the coordination of "planning" and "operation". This method can reduce the investment and operation costs of the active distribution network while improving the reliability, safety and sustainability of the distribution network during operation. In the planning and optimization of distributed power sources, considering load demand response and the access of energy storage equipment can reasonably and effectively absorb distributed power sources, reduce the total economic cost, and promote voltage stability and load balance during the operation of the distribution network. In the process of solving the two-level planning problem, the proposed ICSQPSO algorithm not only has a strong global optimization capability, thereby avoiding the solution from falling into the local optimal solution; it also has higher computing efficiency and saves running time.

[0282] Embodiment 2

[0283] The present invention also provides an electronic device, comprising: a memory and a processor; the memory is used to store a program; the processor is used to execute the program to implement each step of the ADN source-load-storage collaborative double-layer optimization configuration method considering demand response.

[0284] Embodiment 3

[0285] The present invention also provides a readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the ADN source-load-storage collaborative double-layer optimization configuration method considering demand response are implemented.

[0286] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for optimizing the coordinated dual-layer configuration of ADN source-load-storage considering demand response, characterized in that: The following steps are involved: Considering demand response, an ADN source-load-storage model is constructed, wherein the ADN source-load-storage model includes a planning layer and an operation layer; The Metropolis criterion in the SA algorithm is integrated into the quantum-behaved particle swarm optimization algorithm, and a dynamic adjustment strategy based on the Levy distribution is introduced in the Levy flight of the cuckoo search algorithm; The improved quantum behavior particle swarm optimization algorithm and the cuckoo search algorithm are fused to obtain an improved hybrid cuckoo search quantum behavior particle swarm optimization algorithm; Based on the improved hybrid cuckoo search quantum behavior particle swarm optimization algorithm, the variables in the planning layer and the operation layer are solved to complete the configuration.

2. The ADN source-load-storage collaborative dual-layer optimization configuration method considering demand response according to claim 1 is characterized in that: The process of building an ADN source-load-storage model includes: The power supply side is modeled based on the output characteristics of wind power generation, photovoltaic power generation and gas turbine power generation; the load side is modeled based on the rigid load response output characteristics and the flexible load response output characteristics; the energy storage model is modeled based on the state of charge of the battery constraints; and the ADN source-load-storage model is constructed based on the constructed power supply side, load side and energy storage side.

3. The ADN source-load-storage collaborative dual-layer optimization configuration method considering demand response according to claim 1 is characterized in that: The demand response adopts the time-of-use electricity price mechanism in the price-driven demand response, and combines the interruptible load and transferable load in the incentive-driven demand response.

4. The ADN source-load-storage collaborative dual-layer optimization configuration method considering demand response according to claim 1 is characterized in that: The planning layer is used to obtain the total planning layer cost by summing the investment and construction costs of each device and the results returned by the operation layer planning; then, the power balance constraint, the capacity constraint of the newly built generator set and the proportion of new energy access constraints are used as constraints, and the minimization of the annual comprehensive planning operation cost is used as the objective function to obtain the optimal planning scheme for the site selection and capacity determination of each device.

5. The ADN source-load-storage collaborative dual-layer optimization configuration method considering demand response according to claim 4 is characterized in that: The operation layer is used to obtain the total cost of the operation layer by weighted summing the economic cost of the operation layer, the voltage offset of the distribution network node, the wind power abandonment rate, the photovoltaic abandonment rate and the standard deviation of the equivalent load curve; then, the unit output constraint, the interruptible load constraint, the transferable load constraint, the energy storage device output constraint, the upper power purchase constraint, the power flow constraint and the node voltage constraint are used as constraint conditions, and the total cost of the operation layer is used as the objective function to obtain the optimal scheduling result of the optimal planning scheme of the response planning layer, and the optimal scheduling result includes the daily scheduling curve and the current distribution network operation status; Among them, the economic costs at the operating level include the operation and maintenance costs of power supply equipment, the management costs of load demand response, the depreciation costs of energy storage equipment, the annual electricity purchase costs from the superior power grid, and network loss costs.

6. The ADN source-load-storage collaborative dual-layer optimization configuration method considering demand response according to claim 1 is characterized in that: The process of solving the variables in the planning layer and the operation layer based on the improved hybrid cuckoo search quantum behavior particle swarm optimization algorithm includes: Input the parameters of the ADN source-charge-storage model into the improved hybrid cuckoo search quantum behavior particle swarm optimization algorithm to initialize the algorithm parameters; Initialize the location and capacity of the power source and energy storage equipment in the ADN source-load-storage model as the initial decision variables of the planning layer; Inputting the initial decision variables of the planning layer into the operation layer for solving, and obtaining the optimal solution of the operation layer as the decision variables of the operation layer; The decision variables of the operation layer are input into the planning layer for solving, and the initial decision variables of the planning layer are updated to obtain the optimal solution of the planning layer; The optimal solution of the operation layer and the optimal solution of the planning layer are input into the planning layer, and the objective function is calculated as the fitness; The new fitness is compared with the old fitness, and the Metropolis criterion in the SA algorithm is used to determine whether to accept the new solution, so as to update and iterate the optimal solutions of the planning layer and the operation layer until the maximum number of iterations is reached, and then the optimal solutions of the planning layer and the operation layer are output.

7. The ADN source-load-storage coordinated dual-layer optimization configuration method considering demand response according to claim 6 is characterized in that: The optimal solutions at the planning level include: The optimal solutions for the number of wind turbine units per unit capacity to be built, the number of photovoltaic power generation units per unit capacity to be built, and the number of gas turbine units per unit capacity to be built constitute a solution set as the optimal planning result.

8. The ADN source-load-storage coordinated dual-layer optimization configuration method considering demand response according to claim 6 is characterized in that: The optimal solutions for the operation layer include: The optimal solutions for the active output of each wind power station, the active output of each photovoltaic power station, and the active output of each gas turbine power station constitute a solution set as the optimal scheduling result.

9. An electronic device, characterized in that: include: Memory and processor; The memory is used to store programs; The processor is used to execute the program to implement each step of the ADN source-load-storage collaborative dual-layer optimization configuration method considering demand response as described in any one of claims 1-8.

10. A readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, each step of the ADN source-load-storage collaborative dual-layer optimization configuration method considering demand response as described in any one of claims 1 to 8 is implemented.

Images