An energy storage control method for an active distribution network containing new energy with high permeability

Abstract

The present invention relates to a control method for energy storage in an active distribution network containing new energy with high penetration rate, including steps of studying a coordinated control strategy for energy storage to assist new energy units in participating in primary frequency regulation; establishing a coordinated control model for energy storage to assist new energy units in participating in primary frequency regulation; analyzing the power response requirements for energy storage to participate in power grid peak shaving and realizing capacity optimization allocation; proposing an optimized control method for a virtual power plant considering the characteristics of new energy units and energy storage to participate in AGC frequency regulation; the present invention improves the frequency regulation and peak shaving capabilities of the distribution network by coordinating new energy and energy storage devices, thereby enhancing the new energy consumption ratio and economic benefits of the distribution network; for power supply enterprises, improving the new energy consumption capacity is an embodiment of environmental protection and fulfilling social responsibilities; the improvement of key technologies for frequency regulation and peak shaving of energy storage in a distribution network with high penetration rate can maximize the grid connection and consumption of new energy resources within the region, enhance the energy conservation and emission reduction benefits within the region, and improve the end - electrification level.

Description

Technical Field

[0001] The present invention relates to the technical field of new energy distribution network energy storage control, and specifically relates to an energy storage control method for an active distribution network with a high penetration rate of new energy. Background Art

[0002] In recent years, power grids with a high proportion of new energy have frequently experienced large-scale power outages after suffering from faults; mainly because the proportion of new energy in the system is very high, the overall system moment of inertia is insufficient, and it is unable to instantly provide sufficient energy support for the power grid, resulting in an excessive rate of change of system frequency and causing the under-frequency load shedding protection to lock out; this causes a large amount of active power deficit to be completely transferred to the tie lines, and the tie lines are disconnected due to overload, ultimately leading to system frequency collapse; in addition, the originally set maximum disturbance amount of the power grid can resist general lightning strike accidents, but due to the insufficient grid-connected technology of new energy units, new energy units are prone to tripping due to lightning strikes, aggravating the power deficit in the system and causing large-scale power outages; moreover, new energy units will also trip due to triggering frequency protection, aggravating the active power deficit of the power grid and resulting in long-term frequency deviation of the system; the existing new energy power grids mainly have prominent problems with insufficient system frequency regulation ability after suffering from faults and new energy unit accidental tripping; therefore, it is urgent to study practical energy storage efficient control and optimal configuration technologies to meet the efficient consumption of new energy, while improving the frequency regulation ability and operation efficiency of the power grid. Summary of the Invention

[0003] In view of the deficiencies of the prior art, the present invention provides an energy storage control method for an active distribution network with a high penetration rate of new energy, characterized in that: the active distribution network energy storage control method includes the following steps:

[0004] Step 1: Study the coordinated control strategy of energy storage assisting new energy units to participate in primary frequency regulation;

[0005] Step 2: Establish a coordinated control model of energy storage assisting new energy units to participate in primary frequency regulation;

[0006] Step 3: Analyze the power response requirements of energy storage participating in power grid peak shaving and realize capacity optimization configuration;

[0007] Step 4: Propose an optimal control method for a virtual power plant considering the characteristics of new energy units and energy storage to participate in AGC frequency regulation.

[0008] Furthermore, the coordinated control strategy of energy storage assisting new energy units to participate in primary frequency regulation includes: a comprehensive control strategy of energy storage assisting new energy units to participate in system frequency regulation / peak shaving, an energy management strategy of energy storage batteries, and economic analysis.

[0009] Furthermore, the comprehensive control strategy for the energy storage-assisted new energy unit to participate in system frequency regulation / peak shaving is that according to different grid frequencies, the energy storage battery can operate in two working modes: frequency regulation / peak shaving; under these two modes, the control of the photovoltaic-storage system includes the following three main links:

[0010] (a) The photovoltaic system adopts maximum power point tracking control, and the photovoltaic-storage coordinated control system detects the grid frequency and determines whether the frequency is within the frequency regulation dead zone;

[0011] (b) When the grid frequency is within the frequency regulation dead zone, the energy storage battery operates in the peak shaving mode, and at this time, the charging and discharging of the energy storage battery are controlled according to the peak shaving period; when the grid frequency exceeds the frequency regulation dead zone, the energy storage battery operates in the frequency regulation mode, and the primary reference power of the energy storage battery is obtained according to the frequency deviation direction;

[0012] (c) Compare with the idle capacity of the inverter, and take the smaller value of the two as the secondary reference power of the energy storage battery, which is optimized by the maximum output constraint coefficient of the energy storage battery as the final reference power of the energy storage battery.

[0013] Furthermore, the energy management strategy of the energy storage battery is based on the comprehensive control strategy of the photovoltaic-storage system participating in grid frequency regulation / peak shaving, and designs the state of charge partition of the energy storage battery and the energy management strategy of the energy storage battery in different frequency intervals;

[0014] When partitioning the state of charge of the energy storage battery, the priority of the frequency regulation mode is higher than that of the peak shaving mode; to ensure that the energy storage battery leaves a capacity margin for the frequency regulation mode during peak shaving, when partitioning the state of charge of the energy storage battery, two situations must be considered: (1) Overcharging / overdischarging of the energy storage battery will damage the battery. To avoid this situation, the state of charge intervals for charging and discharging the energy storage battery in the frequency regulation mode are the upper limit of charging and the lower limit of discharging. The upper limit of charging is represented by SOC min , and the lower limit of discharging is represented by SOC max . When the charging reaches the upper limit SOC max , charging stops, and when the discharging reaches the lower limit SOC min , discharging stops;

[0015] (2) To ensure that the energy storage battery leaves a capacity margin for the frequency regulation mode in the peak shaving mode, therefore, it is set that in the peak shaving mode of the energy storage battery, the upper limit of charging is SOC high , and the lower limit of discharging is SOC 1ow ; the state of charge partition diagram of the energy storage battery is obtained.

[0016] Furthermore, the economic analysis refers to:

[0017] Analyze the economy of the photovoltaic-storage combined system participating in grid frequency regulation and peak shaving during the entire life cycle of the energy storage battery after adding the energy storage battery to the photovoltaic power station; the increased cost of reconstructing the photovoltaic power station is:

[0018] C = C1 + C2 + C3 - C4 (1)

[0019] Where: C is the total increased cost of the photovoltaic - energy storage combined system; C1 is the initial cost of reconstructing the photovoltaic power station; C2 is the operation and maintenance cost of the energy - storage battery; C3 is the loss cost caused by the efficiency of the energy - storage battery; C4 is the residual value of the power station;

[0020] The energy - storage converter installed during the initial construction can still be used after the end of the battery cycle life. Therefore, the initial cost only considers the capacity cost of the energy - storage battery, that is

[0021]

[0022] Where, k1 is the unit capacity cost; E is the capacity of the energy - storage battery; k2 is the operation and maintenance cost per unit capacity of the energy - storage battery; E′ is the cumulative value of the charge - discharge amount of the energy - storage battery during its entire life cycle, that is

[0023] E′ = E×(SOC max - SOC min )×2n (3)

[0024] Where: n is the number of cycles of the energy - storage battery during its entire life cycle;

[0025] C3 is the loss cost caused by the efficiency of the energy - storage battery, that is:

[0026]

[0027] Where: k3 is the grid - connected electricity price of the photovoltaic power; η is the charge - discharge efficiency of the energy - storage battery;

[0028] C4 is the residual value of the power station, that is:

[0029] C4 = k4×C1 (5)

[0030] Where: k4 is the recovery value coefficient after the energy - storage battery has completed its service life:

[0031] The benefits brought by adding the energy - storage battery are:

[0032] R = R1 + R2 + R3 (6)

[0033] Where: R is the total increased revenue after adding the energy - storage battery; R1 is the compensation revenue obtained by the photovoltaic - energy storage system participating in frequency regulation; R2 is the compensation revenue obtained by the photovoltaic - energy storage system participating in peak regulation; R3 is the additional electricity - selling revenue obtained when all the photovoltaic power output is grid - connected compared with the photovoltaic load - shedding mode;

[0034]

[0035] Where: k5 is the frequency modulation compensation price per unit energy; E″ is the cumulative value of the operation amount of the energy storage battery participating in frequency modulation during its entire life cycle; k6 is the peak shaving compensation price per unit energy; E2′ is the cumulative value of the operation amount of the energy storage battery participating in peak shaving during its entire life cycle;

[0036] Let the cumulative value of the operation amount of the energy storage battery participating in peak shaving and frequency modulation within 24 hours of a typical day be E″, then the number of days in the entire life cycle of the energy storage battery is

[0037]

[0038] Then the increased revenue from all grid-connected photovoltaic power output is

[0039] R3 = E pv × 33% × k3 × t (9)

[0040] Where: E pv is the total energy generated by the photovoltaic during 24 hours in the maximum power point operation mode of a typical day.

[0041] Furthermore, in step 2, a coordinated control model for the energy storage to assist the new energy unit in participating in primary frequency modulation is established; that is, numerical modeling is carried out on the energy storage converter and its control system, and then a simulation model of the active distribution network containing high-penetration new energy is established and a coordinated control model for the new energy unit and the energy storage to participate in frequency modulation is constructed to analyze the dynamic response performance of the energy storage device during photovoltaic power output fluctuations or rapid load fluctuations, and analyze the frequency modulation effect of the energy storage device under different load levels and different operating conditions;

[0042] Mathematical model of the energy storage battery peak shaving mode: In the peak shaving mode, the charging and discharging duration of the energy storage battery is at the hour level. Therefore, the energy storage battery can charge and discharge at a relatively small power according to its own capacity size to extend the service life of the energy storage battery; in actual engineering, when the state of charge constraint is not considered, the primary reference power for charging and discharging in the peak shaving mode is usually set to 0.3 times the rated power of the energy storage battery; according to the instructions of the dispatching center, while ensuring that the energy storage battery is not overcharged / overdischarged, a certain capacity margin is reserved for the frequency modulation mode of the energy storage battery; therefore, the primary reference power and the state of charge constraint of the energy storage battery in the peak shaving mode are set as:

[0043]

[0044] Where: P e ′ ss is the primary reference power of the energy storage battery, and P e is the rated power of the energy storage battery; it is stipulated that the discharge direction is positive;

[0045] Mathematical Model of Energy Storage Battery Frequency Modulation Mode: After the grid frequency exceeds the frequency modulation dead zone, the energy storage battery operates in the frequency modulation mode; the frequency modulation mode requires that the energy storage battery can absorb or emit as much power as possible in a short time to support the grid frequency. Therefore, the energy storage battery is set to charge and discharge at the rated power in the frequency modulation mode; on the premise of ensuring no overcharge / overdischarge, the energy storage battery charges and discharges according to the deviation direction of the grid frequency. The primary reference power and state of charge constraint of the energy storage battery in the frequency modulation mode are set as follows:

[0046]

[0047] In the formula: f is the grid frequency detected by the system, and △f = f - 50;

[0048] Design of the Maximum Output Constraint of Energy Storage Battery Based on SOC Feedback: When the energy storage battery responds to the grid frequency modulation / peak shaving demand, if it charges and discharges at a constant power, it will cause the energy storage battery to not be fully charged or fully discharged, thus wasting the capacity of the energy storage battery and causing economic losses; therefore, a reasonable maximum output constraint coefficient of the energy storage battery should be designed, represented by λ SOC to make the energy storage battery participate in the grid frequency modulation / peak shaving with variable charge and discharge power;

[0049] (a) When the state of charge of the energy storage battery is relatively high, that is, SOC > 50%, it discharges according to P e ′ s ′ s and P e ′ s ′ s is the secondary reference power of the energy storage battery; when discharging to a relatively low state of charge of the energy storage battery, that is, SOC < 50%, to make full use of the capacity of the energy storage battery and avoid overdischarge, the energy storage battery discharges at P e ′ s ′ s multiplied by a λ SOC less than 1, and λ SOC becomes smaller as the state of charge decreases;

[0050] (b) When the state of charge of the energy storage battery is relatively low, that is, SOC < 50%, it charges according to P e ′ s ′ s and when charging to a relatively high state of charge of the energy storage battery, that is, SOC > 50%, to make full use of the capacity of the energy storage battery and avoid overcharge, the energy storage battery charges at P e ′ s ′ s s multiplied by a λ SOC less than 1, and λ SOC decreases as the state of charge increases; in the charge and discharge state, λ SOCAs shown in Equation (12) and Equation (13):

[0051] Under the charging state, there is

[0052]

[0053] Under the discharging state, there is

[0054]

[0055] In the formula: λ SOC is the maximum output constraint coefficient of the energy storage battery, and SOC represents the state of charge of the energy storage battery;

[0056] The constraint relationship between the final reference power of the energy storage battery and the state of charge is:

[0057] P ess = P″ ess ×λ SCC (14)

[0058] In the formula: P ess is the final reference power of the energy storage battery;

[0059] In summary, when the output of the energy storage battery is optimized with the above maximum output constraint coefficient, it can ensure that the energy storage battery has the ability of rapid response, while fully utilizing the capacity of the energy storage battery and avoiding overcharging / overdischarging of the energy storage battery, thus prolonging the service life of the energy storage battery.

[0060] Furthermore, analyzing the power response demand of the energy storage participating in the grid peak regulation and realizing the capacity optimization configuration are achieved by clustering the typical daily load curves and daily new energy generation curves in different seasons to form the "duck" curve of the power grid; selecting typical lines, calculating and analyzing based on the structure of the actual line, new energy installation and user distribution, and establishing a distribution network simulation model;

[0061] According to the ramp rate and minimum output level of conventional units and the power flow limit constraints of tie lines, a capacity optimization configuration method for the energy storage participating in peak regulation based on the particle swarm algorithm is proposed; among them, the peak-valley-flat period division strategy is as follows; the traditional peak-valley-flat period division often only considers the peak-valley-flat membership degree of the load, but considering the large-scale consumption of new energy, it is necessary to comprehensively consider the contribution degree of user load demand and new energy consumption in different time periods to the peak-valley difference of the "duck" curve. Therefore, an improved peak-valley membership function is used to calculate the peak-valley-flat membership degree at each moment, as shown in the following formula:

[0062]

[0063] In the formula: u ft is the peak membership degree of the "duck" curve; u gtis the membership degree of the "duck curve" valley; L(t) is the user load at time t; max(L) and min(L) are the peak and valley values of the load respectively; L c (t) is the new energy output at time t; max(L c ) and min(L c ) are the maximum and minimum values of the new energy output respectively; μ1 and μ2 are the weights of the user load part in the calculation of the peak and valley membership degrees;

[0064] In Equation (15), the larger the peak membership degree at a certain moment, the greater the possibility that it belongs to the peak period; the larger the valley membership degree, the greater the possibility that it belongs to the valley period; therefore, the time period to which the moment belongs is divided by determining the peak-valley membership degree threshold.

[0065]

[0066] In the formula: T f , T p , T g are the peak, flat, and valley periods respectively; m1 is the peak membership degree threshold; m2 is the valley membership degree threshold.

[0067] 1) Optimization model

[0068] Taking into account the frequency modulation effect and economic benefits, optimize the location and capacity configuration of energy storage; the upper-level planning takes the comprehensive optimization of the location and capacity of new energy units and energy storage devices, network loss costs, and frequency fluctuations as the objective function, and the lower-level planning considers the ramping rate, output level, and power flow limit constraints on the switching mode of controllable phase shifters on this basis; the K-means clustering method is used to reduce the scenarios, and the combined harmony search algorithm and particle swarm algorithm are used to jointly solve the model.

[0069] The two-layer programming model is a system optimization model with a two-layer hierarchical structure; the upper-layer model in this embodiment is used to determine the configuration method of reactive power compensation, and the lower-layer model solves the optimal operation mode of the system under the constraint conditions on the basis of the upper-layer model.

[0070] The upper-level planning objective function is:

[0071]

[0072]

[0073] In the formula, n is the number of network branches; U i,t and U j,t are the voltage amplitudes of nodes i and j at time t respectively; G ij , B ij , and ij are the conductance, susceptance, and phase angle difference between nodes i and j respectively.

[0074]

[0075] In the formula, is the switching cost of the shunt capacitor bank, is the operation cost of the power spring; λ SCB and λ ES are the voltage regulation cost weight coefficients of the SCB and ES respectively, and N c is the node where the reactive power compensation device is installed in the network.

[0076]

[0077] In the formula, N is the total number of nodes in the distribution network, and U i is the actual voltage of node i.

[0078] The constraint conditions include:

[0079] Power flow equation constraint

[0080]

[0081] In the formula, n is the number of network branches; P i and Q i are the active power and reactive power injected into each node respectively.

[0082] ② Control variable constraint

[0083] U i,min ≤U i ≤U i,max

[0084]

[0085]

[0086]

[0087]

[0088]

[0089]

[0090] U i,max and U i,min are the upper and lower limit values of the voltage of node i respectively; is the number of reserved capacitors for quasi-steady-state reactive power reserve, and Qmin and Qmax are the minimum and maximum values of the reactive power compensation capacity of the node.

[0091] 2) Model solution

[0092] a. Upper-layer model solution

[0093] The harmony search algorithm is derived from the simulation of achieving the optimal performance effect by harmonizing musical notes in music performance. The harmony search algorithm has the advantages of novel solution method, easy to understand, strong robustness, etc., and is widely used in various fields. The steps of the harmony search algorithm are as follows:

[0094] ① Determine the parameters of the harmony search algorithm;

[0095] ② Initialize the harmony memory library HM;

[0096] ③ Randomly select a new solution from HM with probability HMCR, otherwise randomly select a new solution outside HM with probability 1 - HMCR. If a new solution is selected within HM, the new solution needs to be locally perturbed with probability PAR, and the perturbation variable is bw;

[0097] ④ Update HM;

[0098] ⑤ Judge whether the end condition is satisfied. If satisfied, output the result; if not satisfied, repeat steps ③ and ④;

[0099] The improved harmony search algorithm is adopted, that is, PAR and bw are dynamically adjusted by equations (17) and (18). The improved harmony search algorithm is beneficial to global search in the early stage of iteration and refined search in the later stage of iteration;

[0100] b. Solution of the lower-layer model

[0101] The particle swarm optimization (PSO) algorithm is an intelligent optimization algorithm that simulates the foraging behavior of bird flocks. Each particle in PSO is a candidate solution. PSO consists of three elements: velocity, position, and fitness value. Since the particle swarm is prone to fall into the local optimal solution in the later stage, the particle swarm algorithm combined with the improved weight and immune algorithm is adopted, so that the weight coefficient is changed from a constant to a linearly decreasing one. Inspired by the immune algorithm, in the iterative process of particle swarm solution, the first N better solutions are saved as the optimal values, injected into the population, and the particle types in the population are updated to increase the possibility of the population to find the optimal solution.

[0102] Furthermore, the optimal control method for a virtual power plant with new energy units and energy storage characteristics to participate in AGC frequency regulation is as follows: New energy units such as wind and solar are connected to the grid in a clean and efficient manner, effectively solving the problem of the decreasing traditional fossil energy. As the penetration rate of new energy units increases, the volatility and uncertainty brought about by the imbalance between power generation and load on the source and load sides will lead to the gradual deterioration of wind and solar curtailment, and even cause the instability of system operation; After the energy storage battery is connected to the intelligent distribution network, it has the advantages of peak shaving and valley filling, improving power quality, etc. Reasonably regulating the energy storage capacity can effectively improve the above phenomena. The virtual power plant attracts and aggregates various types of distributed energy through advanced information technology and control technology to participate in power grid dispatching and electricity market transactions. By optimizing the output of distributed energy, it provides a more stable output and can participate in power grid operation as a special power plant; It mainly includes the construction of a wind-solar-storage model and an optimization method based on an improved genetic algorithm.

[0103] Furthermore, the construction of the wind-solar-storage model includes the following:

[0104] a. Wind power output probability model

[0105] The actual output P of the wind turbine w Affected by the change of wind speed, its output probability model can be expressed as:

[0106]

[0107] In the formula: P r represents the rated power of the wind turbine; v pi , v po represent the cut-in and cut-out wind speeds respectively; v r represents the rated wind speed.

[0108] The wind speeds in most areas approximately follow a two-parameter Weibull distribution, and its probability model can be expressed as:

[0109]

[0110] In the formula: v represents the real-time wind speed; k represents the shape parameter; c represents the scale parameter.

[0111] b. Photovoltaic output probability model

[0112] Within a certain time period, the change of light intensity approximately follows a Beta distribution, and its probability model can be expressed as

[0113]

[0114] In the formula: represents the light intensity; φ and δ both represent shape parameters; Γ represents the Gramma function.

[0115] The actual output of photovoltaic (PV) is affected by the change of light intensity, and its output probability model is:

[0116]

[0117] In the formula: Pl(t) represents the output power of light intensity at the t-th moment; G(t) represents the surface temperature of the PV power generation system at the t-th moment; θ represents the temperature coefficient; lste, Gste, and Pste represent the light intensity, the temperature of the PV power generation system, and the maximum output power under the standard test environment, respectively.

[0118] c. Charge and discharge probability model of energy storage battery

[0119] The energy storage is a sodium-sulfur battery with the characteristics of high efficiency and large capacity. From the analysis of its principle, the power difference at the t-th moment can be expressed as

[0120]

[0121] In the formula: ΔP ESB (t) represents the difference between source and load supply and demand at the t-th moment; P load (t) represents the load demand at the t-th moment; PDG_y represents the output power of the y-th type of DG at the t-th moment. If ΔP ESB (t) is greater than 0, the energy storage is in the discharge state; if ΔP ESB (t) is less than 0, the energy storage is in the charging state, and the stored energy at the corresponding t-th moment can be expressed as

[0122]

[0123] In the formula: E ESB (t) represents the stored energy of the energy storage at the t-th moment; ζ represents the charging efficiency of the energy storage, and its value range is [0.7, 0.9].

[0124] d. Optimization scheduling model based on virtual power plant and time-of-use electricity price

[0125] To solve the problems of the volatility of wind and solar power generation and the economy of power grid operation, a two-stage optimization scheduling model is established. The new energy units and energy storage are considered as a virtual power plant. The first stage is the optimization of the output of the virtual power plant. Considering the volatility problems of wind power and PV comprehensively, the energy storage device is used to absorb the fluctuation components, and the output of the virtual power plant after suppressing the fluctuation is transmitted to the second-stage power grid economic optimization stage; the original outputs of PV and wind turbines are known conditions, the minimum fluctuation of the output of the virtual power plant is the objective function, and the charge and discharge of the energy storage battery is used as the control strategy. The most important constraint is the power balance constraint, that is, to meet the active power demand under different active power response requirements:

[0126] P G (t)+PWT (t) + P PV (t) - P ES (t) - P L (t) = 0 (29)

[0127] where: P G (t) is the sum of the active power of all thermal power generating units in the system during period t; P WT (t) is the active power of wind power during period t; P PV (t) is the active power of photovoltaic power generation during period t; P ES (t) is the charge-discharge power of the energy storage device during period t, and when it is positive, it represents charging, P L (t) is the load power; On this premise, an optimization dispatching model based on virtual power plants and time-of-use electricity prices is established, which is divided into two types:

[0128] The first type. Virtual power plant output optimization model

[0129] ① Objective function

[0130] The optimization objective function of the virtual power plant is to minimize the output fluctuation of the virtual power plant:

[0131]

[0132] where: F1 is the volatility index of the output of the virtual power plant; P VPP.t is the active power of the virtual power plant during period t; P VPP.AV is the average value of the active power of the virtual power plant during the entire dispatching cycle. Among them

[0133]

[0134] where: P W (t) is the active power of wind power during period t; P l (t) is the active power of photovoltaic power during period t; E ESB (t) is the charge-discharge active power of the energy storage device.

[0135] ② Constraint conditions

[0136] The constraints of the energy storage device are divided into charge-discharge power limit and remaining capacity limit; The charge-discharge power limit is as follows;

[0137] |E ESB (t)| ≤ E ESB.max (32)

[0138] where: E ESB (t) is the charge-discharge active power of the energy storage device during period t; E ESB.max is the maximum charge-discharge active power of the energy storage device;

[0139] The remaining capacity limit is:

[0140] S min ≤S OC (t) ≤ S max (33)

[0141] where: S oc (t) is the remaining capacity of the energy storage device at time t; S min is the minimum allowable remaining capacity value of the energy storage device; S max is the maximum allowable remaining capacity value of the energy storage device;

[0142] Wind power and PV power output constraints

[0143]

[0144] where: P W (t) is the actual active power of the wind farm at time t; P Wf (t) is the predicted active power of the wind farm at time t. P l (t) is the actual active power of the PV power plant at time t; P l.f (t) is the predicted active power output of the PV power plant at time t.

[0145] Second. Grid economic optimization model

[0146] In the grid economic optimization stage, with the lowest total system operation cost as the objective function, given the optimized output of the virtual power plant, the electricity load curve, and the operating parameters of each thermal power unit, the optimal output of each thermal power unit is obtained;

[0147] ① Objective function

[0148] To achieve the safe and economic operation of each component of the power grid, in the grid economic optimization stage, the lowest operating cost of the entire system is taken as the objective function; as renewable energy sources, wind power and PV do not consume primary energy and have low power generation operating costs, so their power generation costs and environmental costs can be ignored. At the same time, to encourage the development of renewable energy, a certain penalty cost is added for wind and PV curtailment in the system. Time-of-use electricity prices guide users to change their electricity consumption time through electricity prices, and their scheduling costs can be ignored. Therefore, the economic dispatch objective function of the power grid is:

[0149]

[0150] where: F2 is the operating cost of the entire system; C G.t , C SR.t , C new.t , C ES.t are the thermal power generation cost, spinning reserve cost, renewable energy curtailment penalty, and energy storage device depreciation cost in the system at time t, respectively.

[0151] ② Constraint conditions

[0152] The ramp power constraint of the thermal power unit is:

[0153]

[0154] In the formula: P i (t - 1) is the active power of unit i in the (t - 1) period; R up i is the maximum rising power of thermal power unit i in a single period; R dow n i is the maximum falling power of thermal power unit i in a single period.

[0155] Furthermore, the optimization method based on the improved genetic algorithm

[0156] The genetic algorithm originated from the computer simulation research on biological systems. It is a stochastic global search and optimization method developed by imitating the biological evolution mechanism in nature, drawing on Darwin's theory of evolution and Mendel's genetic theory. Its essence is an efficient, parallel, and global search method that can automatically acquire and accumulate knowledge about the search space during the search process and adaptively control the search process to find the optimal solution;

[0157] For the optimization of the output of the virtual power plant in the first stage, the minimum fluctuation is taken as the objective function, and under the condition that the energy storage capacity configuration is determined, its charge and discharge strategy is determined;

[0158] For the second stage, with the economic dispatch economy of the power grid being the optimal objective function, on the premise that the thermal power unit meets the ramp, the maximum output of the new energy unit and the energy storage unit is set as the optimization variable for solution.

[0159] The beneficial effects of the present invention are:

[0160] The present invention solves the problems that due to the high proportion of new energy in the system, the overall system inertia is insufficient, and it is unable to instantaneously provide sufficient energy support for the power grid, resulting in an excessive frequency change rate of the system and the locking of the under - frequency load shedding protection; this causes all the active power shortages to be transferred to the tie - line, and the tie - line is disconnected due to overload, ultimately leading to the collapse of the system frequency; also, the originally set maximum disturbance amount of the power grid can resist general lightning strike accidents, but due to the insufficient grid - connected technology of new energy units, new energy units are prone to tripping due to lightning strikes, aggravating the power shortage in the system and causing large - scale power outages; moreover, new energy units will also trip due to triggering frequency protection, aggravating the active power shortage in the power grid and resulting in long - term frequency deviation of the system; the existing new energy power grid mainly has the problem of insufficient system frequency regulation ability in the accident of unexpected tripping of new energy units after suffering a fault.

[0161] The present invention improves the frequency regulation and peak shaving capabilities of the distribution network by coordinating new energy and energy storage devices, thereby enhancing the new energy consumption ratio and economic benefits of the distribution network; for power supply enterprises, improving the new energy consumption capacity is an embodiment of environmental protection and fulfilling social responsibilities; the improvement of the key technologies for frequency regulation and peak shaving of high-penetration distribution network energy storage can maximize the grid connection and consumption of new energy resources within the region, enhance the energy conservation and emission reduction benefits within the region, actively respond to the national strategic goals of "carbon peak and carbon neutrality", accelerate the transformation of the energy supply side, and improve the end-side electrification level. Description of the Drawings

[0162] Figure 1 It is a comprehensive control strategy diagram for the system frequency regulation / peak shaving of the present invention;

[0163] Figure 2 It is the maximum output constraint coefficient of the energy storage battery of the present invention;

[0164] Figure 3 It is the typical daily load curve of the Dongchengpu 511 line in the embodiment of the present invention;

[0165] Figure 4 It is the calculation flow chart of the hybrid solution algorithm of the present invention;

[0166] Figure 5 It is the flow chart of the optimization method based on the improved genetic algorithm of the present invention. Detailed Embodiments

[0167] Embodiment 1, as Figures 1-5 shown, the present invention provides an energy storage control method for an active distribution network with high-penetration new energy, including researching the coordinated control strategy of energy storage assisting new energy units to participate in primary frequency regulation; establishing a coordinated control model of energy storage assisting new energy units to participate in primary frequency regulation; analyzing the power response requirements of energy storage participating in grid peak shaving and realizing capacity optimization configuration; proposing an optimized control method for a virtual power plant considering the characteristics of new energy units and energy storage to participate in AGC frequency regulation; specifically:

[0168] 1) The comprehensive control strategy of energy storage assisting new energy units to participate in system frequency regulation / peak shaving;

[0169] According to different grid frequencies, the energy storage battery can operate in two working modes: frequency modulation / peak shaving. Under these two modes, the control of the PV energy storage system includes three main links: (a) The PV system adopts maximum power point tracking (MPPT) control. The PV energy storage coordination control system detects the grid frequency and determines whether the frequency is within the frequency modulation dead zone. (b) When the grid frequency is within the frequency modulation dead zone, the energy storage battery operates in the peak shaving mode. At this time, according to the peak shaving period, the charging and discharging of the energy storage battery are controlled. When the grid frequency exceeds the frequency modulation dead zone, the energy storage battery operates in the frequency modulation mode, and the primary reference power of the energy storage battery is obtained according to the frequency deviation direction. (c) Compare with the idle capacity of the inverter, and take the smaller value of the two as the secondary reference power of the energy storage battery. After optimizing by the maximum output constraint coefficient of the energy storage battery, it is used as the final reference power of the energy storage battery.

[0170] 2) Energy management strategy of the energy storage battery

[0171] Based on the comprehensive control strategy of the PV energy storage system participating in grid frequency modulation / peak shaving, the state of charge partition of the energy storage battery and the energy management strategy of the energy storage battery in different frequency intervals are designed.

[0172] State of charge partition of the energy storage battery: In the control strategy proposed in this paper, the priority of the frequency modulation mode is higher than that of the peak shaving mode. To ensure that the energy storage battery leaves a capacity margin for the frequency modulation mode during peak shaving, when partitioning the state of charge of the energy storage battery, two situations need to be considered: (1) Overcharging / overdischarging of the energy storage battery will damage the battery. To avoid this situation, the state of charge interval for charging and discharging the energy storage battery in the frequency modulation mode is (SOC min , SOC max ). When charging reaches the upper limit SOC max , charging stops. When discharging reaches the lower limit SOC min , discharging stops. (2) To ensure that the energy storage battery leaves a capacity margin for the frequency modulation mode in the peak shaving mode, the upper limit of charging is set to SOC high and the lower limit of discharging is set to SOC 1ow for the energy storage battery in the peak shaving mode.

[0173] 3) Economic analysis

[0174] Analyze the economy of the PV energy storage combined system participating in grid frequency modulation and peak shaving during the entire life cycle of the energy storage battery after adding the energy storage battery to the PV power station. The increased cost of reconstructing the PV power station is

[0175] C = C1 + C2 + C3 - C4 (1)

[0176] Where: C is the total increased cost of the photovoltaic - energy storage combined system; C1 is the initial cost of reconstructing the photovoltaic power station; C2 is the operation and maintenance cost of the energy storage battery; C3 is the loss cost caused by the efficiency of the energy storage battery; C4 is the residual value of the power station.

[0177] The energy storage converter installed during the initial construction can still be used after the end of the battery cycle life. Therefore, the initial cost only considers the capacity cost of the energy storage battery, that is

[0178]

[0179] Where: k1 is the unit capacity cost; E is the capacity of the energy storage battery; k2 is the operation and maintenance cost per unit capacity of the energy storage battery; E′ is the cumulative value of the charge - discharge amount of the energy storage battery during its entire life cycle, that is

[0180] E′ = E×(SOC max -SOC min )×2n (3)

[0181] Where: n is the number of cycles of the energy storage battery during its entire life cycle (one cycle includes one charge and one discharge). C3 is the loss cost caused by the efficiency of the energy storage battery, that is

[0182]

[0183] Where: k3 is the grid - connected electricity price of the photovoltaic power; η is the charge - discharge efficiency of the energy storage battery.

[0184] C4 is the residual value of the power station, that is

[0185] C4 = k4×C1 (5)

[0186] Where: k4 is the recovery value coefficient after the energy storage battery has completed its service life.

[0187] The benefits brought by adding the energy storage battery are

[0188] R = R1 + R2 + R3 (6)

[0189] Where: R is the total increased benefit after adding the energy storage battery; R1 is the compensation benefit obtained by the photovoltaic - energy storage system participating in frequency regulation; R2 is the compensation benefit obtained by the photovoltaic - energy storage system participating in peak shaving; R3 is the additional electricity sales benefit obtained when all the photovoltaic power output is grid - connected compared with the photovoltaic load - shedding mode.

[0190]

[0191] Where: k5 is the compensation price per unit energy for frequency regulation; E″ is the cumulative value of the action amount of the energy storage battery participating in frequency regulation during its entire life cycle; k6 is the compensation price per unit energy for peak shaving; E2′ is the cumulative value of the action amount of the energy storage battery participating in peak shaving during its entire life cycle.

[0192] In the example, the frequency regulation power of the photovoltaic-storage combined system is 10 MW. If the photovoltaic power station participates in system frequency regulation by load shedding, the photovoltaic power station should reserve 10 MW of photovoltaic output. However, considering the revenue of the photovoltaic power station, the load shedding rate of the photovoltaic power station is set at 33%. Since the frequency regulation capacity of the photovoltaic power station is related to the photovoltaic output, that is, when the photovoltaic reaches full power generation, the photovoltaic power station has a frequency regulation capacity of 10 MW. In the photovoltaic-storage combined system, the photovoltaic always operates at the maximum power point, and all the electric energy generated by the photovoltaic is finally grid-connected except for the losses.

[0193] Let the cumulative value of the peak shaving and frequency regulation action amounts of the typical-day energy storage battery within 24 hours, and the number of days in the full life cycle of the energy storage battery be E″, then the number of days in the full life cycle of the energy storage battery is

[0194]

[0195] Then the increased revenue from the full grid connection of the photovoltaic output is

[0196] R3 = E pv × 33% × k3 × t (9)

[0197] In the formula: E pv is the total energy generated by the photovoltaic on the maximum power point operation mode within 24 hours of the typical day.

[0198] (2) Establish a coordinated control model for the energy storage to assist the new energy unit to participate in primary frequency regulation

[0199] Numerically model the energy storage converter and its control system, and then establish an active distribution network simulation model with high-penetration new energy and construct a coordinated control model for the new energy unit and the energy storage to participate in frequency regulation. Analyze the dynamic response performance of the energy storage device when the photovoltaic output fluctuates or the load fluctuates rapidly; analyze the frequency regulation effect of the energy storage device under different load levels and different operating conditions (including day and night conditions).

[0200] Mathematical model of the energy storage battery peak shaving mode: In the peak shaving mode, the charging and discharging duration of the energy storage battery is at the hour level. Therefore, the energy storage battery can charge and discharge at a relatively low power according to its own capacity to extend the service life of the energy storage battery. In actual engineering, when the state of charge constraint is not considered, usually 0.3 times the rated power of the energy storage battery is set as the primary reference power for charging and discharging in the peak shaving mode. According to the instructions of the dispatching center, while ensuring that the energy storage battery is not overcharged / overdischarged, a certain capacity margin is reserved for the frequency regulation mode of the energy storage battery. Therefore, the primary reference power and the state of charge constraint of the energy storage battery in the peak shaving mode are set as:

[0201]

[0202] In the formula: P e ′ ssis the primary reference power of the energy storage battery, P e is the rated power of the energy storage battery; it is stipulated that the discharge is in the positive direction.

[0203] Mathematical model of the energy storage battery frequency modulation mode: After the grid frequency exceeds the frequency modulation dead zone, the energy storage battery operates in the frequency modulation mode. The frequency modulation mode requires that the energy storage battery can absorb or emit as much power as possible in a short time to support the grid frequency. Therefore, the energy storage battery is set to charge and discharge at the rated power in the frequency modulation mode. On the premise of ensuring no overcharge / overdischarge, the energy storage battery charges and discharges according to the deviation direction of the grid frequency. The primary reference power and state of charge constraint of the energy storage battery in the frequency modulation mode are set as (f is the grid frequency detected by the system, △f = f - 50):

[0204]

[0205] Design of the maximum output power constraint of the energy storage battery based on SOC feedback: When the energy storage battery responds to the grid frequency modulation / peak shaving demand, if it charges and discharges at a constant power, it will cause the energy storage battery to not be fully charged or fully discharged, thus wasting the capacity of the energy storage battery and causing economic losses. Therefore, a reasonable maximum output power constraint coefficient λ SOC (λ SOC is the maximum output power constraint coefficient of the energy storage battery), so that the energy storage battery participates in the grid frequency modulation / peak shaving with variable charge and discharge power. (a) When the state of charge of the energy storage battery is relatively high (SOC > 50%), it discharges according to P e ′ s ′ s (P e ′ s ′ s is the secondary reference power of the energy storage battery); when discharging to a relatively low state of charge of the energy storage battery (SOC < 50%), to make full use of the capacity of the energy storage battery and avoid over-discharge, the energy storage battery discharges with P e ′ s ′ s multiplied by a λ less than 1 SOC , and λ SOC becomes smaller as the state of charge decreases. (b) When the state of charge of the energy storage battery is relatively low (SOC < 50%), it charges according to P e ′ s ′ s ; when charging to a relatively high state of charge of the energy storage battery (SOC > 50%), to make full use of the capacity of the energy storage battery and avoid overcharge, the energy storage battery charges with P e ′ s ′ s s multiplied by a λ less than 1 SOC , and λ SOC decreases as the state of charge increases. In the charge and discharge state, λSOC As shown in Formula (12), Formula (13) and Figure 2 shown below:

[0206] During the charging state, there is

[0207]

[0208] During the discharging state, there is

[0209]

[0210] The constraint relationship between the final reference power of the energy storage battery and the state of charge is (P ess (where P is the final reference power of the energy storage battery)

[0211] P ess = P″ ess × λ SCC (14)

[0212] In summary, when the output of the energy storage battery is optimized with the above maximum output constraint coefficient, it can ensure that the energy storage battery has the ability of rapid response, while making full use of the capacity of the energy storage battery and avoiding overcharging / overdischarging of the energy storage battery, thus prolonging the service life of the energy storage battery.

[0213] (3) Analyze the power response requirements for the energy storage to participate in grid peak shaving and achieve capacity optimization configuration

[0214] By clustering, obtain the typical daily load curves and daily new energy generation curves in different seasons to form the "duck curve" of the power grid; select typical lines, calculate and analyze based on the structure of the actual lines, new energy installed capacity and user distribution, and establish a distribution network simulation model;

[0215] By analyzing the daily load curves in summer (August 4th), winter (December 27th) and the lowest load level (July 11th), according to the new energy output data, consumption data and load data, derive the "duck curve" and determine the initial peak-valley-flat periods. Among them, the load level of Dongchengpu 511 line is as Figure 3 ;

[0216] Based on the ramp rate and minimum output level of conventional units and the power flow limit constraints of tie lines, propose a capacity optimization configuration method for the energy storage to participate in peak shaving based on the particle swarm optimization algorithm; among them, the peak-valley-flat period division strategy is as follows; the traditional peak-valley-flat period division often only considers the peak-valley-flat membership degree of the load, but considering the large-scale consumption of new energy, it is necessary to comprehensively consider the contribution degree of user load demand and new energy consumption in different time periods to the peak-valley difference of the "duck curve". Therefore, an improved peak-valley membership function is used to calculate the peak-valley-flat membership degree at each moment, as shown in the following formula:

[0217]

[0218] Where: u ft is the membership degree of the "duck curve" peak; u gt is the membership degree of the "duck curve" valley; L(t) is the user load at time t; max(L) and min(L) are the peak and valley values of the load respectively; L c (t) is the new energy output at time t; max(L c ) and min(L c ) are the maximum and minimum values of the new energy output respectively; μ1 and μ2 are the weights of the user load part in the calculation of the peak and valley membership degrees.

[0219] In Equation (15), the larger the membership degree of the peak at a certain moment, the greater the possibility that it belongs to the peak period; the larger the membership degree of the valley, the greater the possibility that it belongs to the valley period. Therefore, the time period to which the moment belongs is divided by determining the peak-valley membership degree threshold.

[0220]

[0221] Where: T f , T p , T g are the peak, flat, and valley periods respectively; m1 is the peak membership degree threshold; m2 is the valley membership degree threshold.

[0222] 1) Optimization model

[0223] Taking into account the frequency modulation effect and economic benefits, optimize the location and capacity configuration of energy storage. The upper-level planning takes the comprehensive optimization of the location and capacity of new energy units and energy storage devices, network loss costs, and frequency fluctuations as the objective function, and the lower-level planning considers the ramp rate, output level, and the switching mode of the controllable phase shifter subject to power flow constraints on this basis; in this embodiment, the K-means clustering method is used to reduce the scenarios, and the harmony search algorithm and particle swarm algorithm are combined to jointly solve the model.

[0224] The two-layer programming model is a system optimization model with a two-layer hierarchical structure; the upper-level model is used to determine the configuration method of reactive power compensation, and the lower-level model solves the optimal operation mode of the system that meets the constraint conditions on the basis of the upper-level model.

[0225] The upper-level planning objective function is:

[0226]

[0227] Where, n is the number of network branches; U i,t and U j,t are the voltage amplitudes of nodes i and j at time t respectively; G ij , B ij , ij are the conductance, susceptance, and phase angle difference between nodes i and j respectively.

[0228]

[0229] Wherein, is the switching cost of the shunt capacitor bank, is the operation cost of the power spring; λ SCB , λ ES are the voltage regulation cost weight coefficients of SCB and ES respectively, and N c is the node where the reactive power compensation device is installed in the network.

[0230]

[0231] Wherein, N is the total number of nodes in the distribution network, and U i is the actual voltage of node i.

[0232] The constraint conditions include:

[0233] ① Power flow equation constraint

[0234]

[0235] Wherein, n is the number of network branches; P i , Q i are the active power and reactive power injected into each node respectively.

[0236] ② Control variable constraint

[0237]

[0238] U i,max and U i,min are the upper and lower limit values of the voltage of node i respectively; is the number of reserved capacitors for quasi-steady-state reactive power reserve, and Q min , Q max are the minimum and maximum values of the reactive power compensation capacity of the node.

[0239] 2) Model solution

[0240] a. Upper-layer model solution

[0241] The harmony search algorithm is derived from the simulation of achieving the optimal performance effect by harmonizing notes in music performance. The harmony search algorithm has the advantages of novel solution method, easy to understand, strong robustness, etc., and is widely used in various fields. The steps of the harmony search algorithm are as follows:

[0242] ① Determine the parameters of the harmony search algorithm.

[0243] ② Initialize the harmony memory library HM.

[0244] ③Randomly select a new solution from HM with probability HMCR, otherwise randomly select a new solution outside HM with probability 1 - HMCR. If a new solution is selected within HM, the new solution needs to be locally perturbed with probability PAR, and the perturbation variable is bw.

[0245] ④Update HM.

[0246] ⑤Judge whether the end condition is satisfied. If it is satisfied, output the result; if not, repeat steps ③ and ④.

[0247] The improved harmony search algorithm is adopted, that is, PAR and bw are dynamically adjusted by equations (17) and (18). The improved harmony search algorithm is beneficial to global search in the early stage of iteration and refined search in the later stage of iteration.

[0248] b. Solution of the lower-layer model

[0249] The particle swarm optimization (PSO) algorithm is an intelligent optimization algorithm that simulates the foraging behavior of bird flocks. Each particle in PSO is a candidate solution. PSO consists of three elements: velocity, position, and fitness value. Since the particle swarm is prone to falling into a local optimal solution in the later stage, a particle swarm algorithm combined with an improved weight and immune algorithm is adopted, making the weight coefficient change from a constant to a linearly decreasing one. Inspired by the immune algorithm, in the iterative process of particle swarm solution, the top N better solutions are saved as the optimal values, injected into the population, and the particle types in the population are updated, increasing the possibility of the population to find the optimal solution; the flow chart of the double-layer model solution is shown in Figure 4 .

[0250] (4) Propose an optimal control method for virtual power plants with new energy units and energy storage characteristics to participate in AGC frequency modulation

[0251] New energy units such as wind and light are connected to the grid cleanly and efficiently, effectively solving the problem of the increasing shortage of traditional fossil energy. With the increase in the penetration rate of new energy units, the volatility and uncertainty brought about by the imbalance between the source and load sides will lead to the gradual deterioration of wind and light abandonment, and even cause the instability of system operation; after the energy storage battery is connected to the intelligent distribution network, it has the advantages of peak shaving and valley filling, improving power quality, etc. Reasonably regulating the energy storage capacity can effectively improve the above phenomena. Through advanced information technology and control technology, the virtual power plant attracts and aggregates various types of distributed energy to participate in power grid dispatching and electricity market transactions. By optimizing the output of distributed energy, it provides a more stable output and can participate in power grid operation as a special power plant.

[0252] (1) Construction of the wind-solar-storage model

[0253] a. Probability model of wind power output

[0254] The actual output Pw of the fan is affected by the change of wind speed, and its output probability model can be expressed as:

[0255]

[0256] Where: P r represents the rated power of the fan; v pi , v po represent the cut-in and cut-out wind speeds respectively; v r represents the rated wind speed.

[0257] In most areas, the wind speed approximately follows a two-parameter Weibull distribution, and its probability model can be expressed as:

[0258]

[0259] Where: v represents the real-time wind speed; k represents the shape parameter; c represents the scale parameter.

[0260] b. Photovoltaic output probability model

[0261] Within a certain period of time, the change in light intensity approximately follows a Beta distribution, and its probability model can be expressed as

[0262]

[0263] Where: represents the light intensity; φ and δ both represent the shape parameters; Γ represents the Gramma function.

[0264] The actual output of photovoltaic (PV) is affected by the change in light intensity, and its output probability model is

[0265]

[0266] Where: P l (t) represents the output power of the light intensity at the t-th moment; G(t) represents the surface temperature of the PV power generation system at the t-th moment; θ represents the temperature coefficient; l ste , G ste and P ste represent the light intensity, the temperature of the PV power generation system, and the maximum output power under the standard test environment respectively.

[0267] c. Charge and discharge probability model of energy storage battery

[0268] The energy storage is a sodium-sulfur battery with characteristics such as high efficiency and large capacity. From the analysis of its principle, the power difference at the t-th moment can be expressed as:

[0269]

[0270] Where: ΔP ESB (t) represents the source-load supply-demand difference at the t-th moment; P load(t) represents the load demand at the t-th moment; P DG_y represents the output power of the y-th type of DG at the t-th moment. If ΔP ESB (t) is greater than 0, the energy storage is in the discharging state; if ΔP ESB (t) is less than 0, the energy storage is in the charging state, and the stored energy at the t-th moment can be expressed as

[0271]

[0272] In the formula: E ESB (t) represents the stored energy of the energy storage at the t-th moment; ζ represents the charging efficiency of the energy storage, and its value range is [0.7, 0.9].

[0273] (2) Optimization scheduling model based on virtual power plant and time-of-use electricity price

[0274] To solve the problems of the volatility of wind and solar power generation and the economy of power grid operation, a two-stage optimization scheduling model is established. The new energy units and energy storage are considered as a virtual power plant. The first stage is the optimization of the output of the virtual power plant. Considering the volatility of wind power and photovoltaic power comprehensively, the energy storage device is used to absorb the fluctuating component, and the output of the virtual power plant after suppressing the fluctuation is transmitted to the second-stage power grid economic optimization stage. The original outputs of photovoltaic and wind turbines are known conditions, the minimum fluctuation of the output of the virtual power plant is the objective function, and the charge and discharge of the energy storage battery is used as the control strategy. The most important constraint is the power balance constraint, that is, the active power demand is satisfied under different active power response demands:

[0275] P G (t)+P WT (t)+P PV (t)-P ES (t)-P L (t) = 0 (29)

[0276] In the formula: P G (t) is the sum of the active powers of all thermal power generating units in the system at the t-th time period; P WT (t) is the active power of wind power at the t-th time period; P PV (t) is the active power of photovoltaic power generation at the t-th time period; P ES (t) is the charge and discharge power of the energy storage device at the t-th time period, and it represents charging when it is positive, P L (t) is the load power.

[0277] On this premise, an optimization scheduling model based on virtual power plant and time-of-use electricity price is established

[0278] a. Virtual power plant output optimization model

[0279] ① Objective function

[0280] The optimization objective function of the virtual power plant is to minimize the output fluctuation of the virtual power plant:

[0281]

[0282] In the formula: F1 is the volatility index of the virtual power plant output; P VPP.t is the active power of the virtual power plant at time t; P VPP.AV is the average value of the active power of the virtual power plant during the entire scheduling period. Among them

[0283]

[0284] In the formula: P W (t) is the active power of wind power at time t; P l (t) is the active power of photovoltaic power at time t; E ESB (t) is the active power of charge and discharge of energy storage devices.

[0285] ② Constraint conditions

[0286] The constraints of energy storage devices are divided into charge and discharge power limits and remaining capacity limits; the charge and discharge power limits are as follows.

[0287] |E ESB (t)| ≤ E ESB.max (32)

[0288] In the formula: E ESB (t) is the active power of charge and discharge of the energy storage device at time t; E ESB.max is the maximum active power of charge and discharge of the energy storage device.

[0289] The remaining capacity limit is:

[0290] S min ≤ S OC (t) ≤ S max (33)

[0291] In the formula: S oc (t) is the remaining capacity of the energy storage device at time t; S min is the minimum allowable remaining capacity value of the energy storage device; S max is the maximum allowable remaining capacity value of the energy storage device.

[0292] Wind power and photovoltaic power output constraints

[0293]

[0294] In the formula: P W (t) is the actual active power of the wind farm at time t; P wF (t) is the predicted active power of the wind farm at time t. P lP(t) is the actual active power of the PV power plant at time t; P l.F P(t) is the predicted active power output of the PV power plant at time t.

[0295] b. Grid economic optimization model

[0296] In the grid economic optimization stage, with the minimum total system operation cost as the objective function, given the optimized output of the virtual power plant, the electricity load curve, and the operating parameters of each thermal power unit, the optimal output of each thermal power unit is obtained.

[0297] ① Objective function

[0298] To achieve the safe and economic operation of each component of the power grid, the grid economic optimization stage takes the minimum operation cost of the entire system as the objective function. As renewable energy sources, wind power and PV power do not consume primary energy and have relatively low power generation operation costs, so their power generation costs and environmental costs can be ignored. At the same time, to encourage the development of renewable energy, a certain penalty cost is added for the curtailment of wind power and PV power in the system; time-of-use electricity prices guide users to change their electricity consumption time through electricity prices, and their dispatching costs can be ignored. Therefore, the economic dispatching objective function of the power grid is

[0299]

[0300] In the formula: F2 is the operation cost of the entire system; C G.t , C SR.t , C new.t , C ES.T are respectively the thermal power generation cost, spinning reserve cost, renewable energy curtailment penalty, and energy storage device depreciation cost in the system at time t.

[0301] ② Constraint conditions

[0302] The ramp power constraint of the thermal power unit is:

[0303]

[0304] In the formula: P i (t - 1) is the active power of unit i at time t - 1; R up i is the maximum upward power of thermal power unit i in a single time period; R down i is the maximum downward power of thermal power unit i in a single time period.

[0305] (3) Optimization method based on improved genetic algorithm

[0306] The genetic algorithm originated from the computer simulation research on biological systems; it is a stochastic global search and optimization method developed by mimicking the biological evolution mechanism in nature, drawing on Darwin's theory of evolution and Mendel's genetic theory. Its essence is an efficient, parallel, and global search method that can automatically acquire and accumulate knowledge about the search space during the search process and adaptively control the search process to find the optimal solution.

[0307] For the optimization of the virtual power plant output in the first stage, the minimum fluctuation is taken as the objective function, and the charge and discharge strategy is determined under the condition that the energy storage capacity configuration is determined.

[0308] For the second stage, the economic optimal operation of the power grid is taken as the objective function, and on the premise that the thermal power unit meets the ramp rate, the maximum output of the new energy unit and the energy storage unit is set as the optimization variable for solution; its algorithm flow is as Figure 5 shown.

[0309] In the specific embodiments described above, the purpose, technical solutions, and beneficial effects of the invention have been further detailed. It should be understood that the above are only specific embodiments of the present invention and are not used to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for controlling energy storage in an active distribution network containing high-penetration new energy, characterized in that: The active distribution network energy storage control method comprises the following steps: Step 1: Study the coordinated control strategy of energy storage-assisted new energy units participating in primary frequency regulation; Step 2: Establish a coordinated control model for energy storage to assist new energy units in participating in primary frequency regulation; Step 3: Analyze the power response requirements of energy storage for grid peak regulation and achieve optimal capacity configuration; Step 4: Propose an optimization control method for a virtual power plant with new energy units and energy storage characteristics to participate in AGC frequency modulation, including wind and solar energy storage model construction and an optimization method based on an improved genetic algorithm. The wind and solar energy storage model construction includes the following: a. Wind power output probability model Actual output of fan P w Affected by wind speed changes, the output probability model is expressed as: Where: P r Indicates the rated power of the fan; v pi 、v po Respectively represent the cut-in and cut-out wind speeds; v r Indicates rated wind speed; In most areas, wind speed approximately obeys the two-parameter Weibull distribution, and its probability model is expressed as: Where: v represents the real-time wind speed; k represents the shape parameter; c represents the scale parameter; b. Photovoltaic output probability model In a certain period of time, the change of light intensity approximately follows the Beta distribution, and its probability model is expressed as: Where: l represents the light intensity; φ and δ represent shape parameters; Γ represents the Gramma function; The actual photovoltaic output is affected by changes in light intensity, and its output probability model is: Where: P l (t) represents the output power of the light intensity at the tth moment; G(t) represents P V The surface temperature of the power generation system at time t; θ represents the temperature coefficient; l ste , G ste and P ste Respectively represent the light intensity under standard test environment, P V Power generation system temperature and maximum output power; c. Energy storage battery charging and discharging probability model: The power difference at time t is expressed as Where: ΔP ESB (t) represents the difference between source and load supply and demand at time t; P load (t) represents the load demand at the tth moment; P DG_y represents the output power of the y-th type DG at the t-th time; if ΔP ESB (t) is greater than 0, the energy storage is in the discharge state; if ΔP ESB (t) is less than 0, the energy storage is in the charging state, and the corresponding stored energy at the tth moment is expressed as Where: E ESB (t) represents the storage energy at the tth moment; ζ represents the energy storage charging efficiency, and its value range is [0.7, 0.9]; d. Optimal scheduling model based on virtual power plant and time-of-use electricity price A two-stage optimization dispatch model is established, and new energy units and energy storage are considered as virtual power plants. The first stage is the optimization of virtual power plant output. The fluctuation component is absorbed by the energy storage device, and the smoothed virtual power plant output is transferred to the second stage of grid economic optimization. The original output of photovoltaic and wind turbines is a known condition, and the minimum fluctuation of virtual power plant output is the objective function. The charging and discharging of energy storage batteries is used as a control strategy to meet the active power demand under different active power response requirements: P G (t)+P WT (t)+P PV (t)-P ES (t)-P L (t)=0 Where: P G (t) is the sum of the active power of all thermal power generating units in the system during the period t; P WT (t) is the active power of wind power in period t; P PV (t) is the active power of photovoltaic power generation in period t; P ES (t) is the charge and discharge power of the energy storage device in time period t, and a positive value indicates charging, P L (t) is the load power; Establish an optimal scheduling model based on virtual power plants and time-of-use electricity prices.

2. The method for controlling energy storage of an active distribution network containing high-penetration new energy according to claim 1, characterized in that: The coordinated control strategy for energy storage-assisted new energy units to participate in primary frequency regulation in step 1 includes: a comprehensive control strategy for energy storage-assisted new energy units to participate in system frequency regulation or peak regulation, energy storage battery energy management strategy, and economic analysis.

3. The method for controlling energy storage of an active distribution network containing high-penetration new energy according to claim 2, characterized in that: The comprehensive control strategy of energy storage assisting new energy units to participate in system frequency regulation or peak regulation is that the energy storage battery can work in two working modes, frequency regulation or peak regulation, according to different grid frequencies; in these two modes, the control of the photovoltaic storage system includes the following three main links: (a) The photovoltaic system adopts maximum power tracking control, and the photovoltaic storage coordinated control system detects the grid frequency and determines whether the frequency is in the frequency regulation dead zone; (b) When the grid frequency is within the frequency regulation dead zone, the energy storage battery operates in the peak regulation mode. At this time, the charging and discharging of the energy storage battery is controlled according to the peak regulation period. When the grid frequency exceeds the frequency regulation dead zone, the energy storage battery operates in the frequency regulation mode, and the primary reference power of the energy storage battery is obtained according to the frequency offset direction. (c) Compared with the idle capacity of the inverter, the smaller value of the two is taken as the secondary reference power of the energy storage battery, and after the maximum output constraint coefficient of the energy storage battery is optimized, it is used as the final reference power of the energy storage battery.

4. The method for controlling energy storage of an active distribution network containing high-penetration new energy according to claim 2, characterized in that: The energy management strategy of the energy storage battery is based on the comprehensive control strategy of the photovoltaic storage system participating in the frequency regulation or peak regulation of the power grid, and designs the charge state partition of the energy storage battery and the energy management strategy of the energy storage battery in different frequency intervals; When the state of charge of the energy storage battery is divided into zones, the priority of the frequency modulation mode is higher than that of the peak regulation mode. To ensure that the energy storage battery has a capacity margin for the frequency modulation mode during the peak regulation process, two situations must be considered when dividing the state of charge of the energy storage battery: (1) Overcharging or over-discharging the energy storage battery will damage the battery. To avoid this, the state of charge range of the energy storage battery in the frequency modulation mode is the upper limit of charging and the lower limit of discharging. The upper limit of charging is expressed as SOC min Indicates that the discharge reaches the lower limit with SOC max Indicates that when the charge reaches the upper limit SOC max When the discharge reaches the lower limit SOC, it will no longer charge. min No more discharge; (2) To ensure that the energy storage battery has enough capacity margin for the frequency modulation mode in the peak modulation mode, the upper limit of the charge of the energy storage battery in the peak modulation mode is set to SOC high , the lower limit of discharge is SOC 1ow ; Obtain the state of charge partition diagram of the energy storage battery.

5. The method for controlling energy storage of an active distribution network containing high-penetration new energy according to claim 2, characterized in that: The economic analysis refers to: The economic benefits of the photovoltaic power station joining the energy storage battery in the full life cycle of the energy storage battery in participating in the frequency and peak regulation of the power grid are analyzed; the additional cost of rebuilding the photovoltaic power station is: C=C1+C2+C3-C4 In the formula: C is the total cost of the photovoltaic and energy storage combined system; C1 is the initial cost of rebuilding the photovoltaic power station; C2 is the operation and maintenance cost of the energy storage battery; C3 is the loss cost caused by the efficiency of the energy storage battery; C4 is the residual value of the power station; The energy storage converter installed during the initial construction can still be used after the battery cycle life ends, so the initial cost only considers the capacity cost of the energy storage battery, that is, In the formula, k1 is the unit capacity cost; E is the energy storage battery capacity; k2 is the unit capacity energy storage battery operation and maintenance cost; E′ is the cumulative value of the charge and discharge amount of the energy storage battery during its entire life cycle, that is, E′=E×(SOC max -SOC min )×2n Where: n is the number of cycles in the entire life cycle of the energy storage battery; C3 is the loss cost caused by the efficiency of the energy storage battery, that is: In the formula: k3 is the photovoltaic grid-connected electricity price; η is the charging and discharging efficiency of the energy storage battery; C4 is the residual value of the power station, that is: C4=k4×C1 Where: k4 is the recovery value coefficient of the energy storage battery after service: The benefits of adding energy storage batteries are: R=R1+R2+R3 Where: R is the total revenue increased after adding energy storage batteries; R1 is the compensation revenue obtained by the photovoltaic storage system participating in frequency regulation; R2 is the compensation revenue obtained by the photovoltaic storage system participating in peak regulation; R3 is the additional electricity sales revenue obtained when all photovoltaic output is connected to the grid compared with the photovoltaic load reduction mode; In the formula: k5 is the compensation price per unit energy for frequency regulation; E1′ is the cumulative value of the amount of action participating in frequency regulation during the entire life cycle of the energy storage battery; k6 is the compensation price per unit energy for peak regulation; E2′ is the cumulative value of the amount of action participating in peak regulation during the entire life cycle of the energy storage battery; Let the cumulative value of the peak load and frequency regulation actions of the energy storage battery in a typical day within 24 hours be E″, then the number of days in the full life cycle of the energy storage battery is: The additional revenue from connecting all photovoltaic power to the grid is: R3=E pv ×33%×k3×t Where: Epv is the total energy emitted by photovoltaic power generation in the maximum power point operation mode for 24 hours on a typical day.

6. The method for controlling energy storage of an active distribution network containing high-penetration new energy according to claim 1, characterized in that: The step 2 establishes a coordinated control model for energy storage-assisted new energy units to participate in primary frequency modulation, which is to perform numerical modeling on the energy storage converter and its control system, and then establish an active distribution network simulation model containing high-penetration new energy and a coordinated control model for new energy units and energy storage to participate in frequency modulation, analyze the dynamic response performance of the energy storage device when the photovoltaic output fluctuates or the load fluctuates rapidly, and analyze the frequency modulation effect of the energy storage device under different load levels and different operating conditions; Mathematical model of peak-shaving mode of energy storage battery: According to the instructions of the dispatching center, a certain capacity margin is reserved for the frequency regulation mode of the energy storage battery under the premise of ensuring that the energy storage battery is not overcharged or over-discharged; therefore, the primary reference power and charge state constraints of the energy storage battery in the peak-shaving mode are set as follows: Where: P′ ess is the primary reference power of the energy storage battery, P e is the rated power of the energy storage battery; the discharge is specified as the positive direction; Mathematical model of frequency modulation mode of energy storage battery: After the grid frequency exceeds the frequency modulation dead zone, the energy storage battery operates in the frequency modulation mode; the frequency modulation mode requires the energy storage battery to absorb or emit more power in a short time to support the grid frequency. Therefore, the energy storage battery is set to charge and discharge at rated power in the frequency modulation mode; on the premise of not overcharging or overdischarging, the energy storage battery is charged and discharged according to the offset direction of the grid frequency, and the primary reference power and charge state constraints of the energy storage battery in the frequency modulation mode are set as follows: Where: f is the grid frequency detected by the system, △f = f-50; Design of maximum output constraint of energy storage battery based on SOC feedback: (a) When SOC>50%, according to P″ ess Discharge, P″ ess is the secondary reference power of the energy storage battery; When SOC<50%, the energy storage battery is P″ ess Multiply by a lambda less than 1 SOC Discharge is performed, and λ SOC It gets smaller as the state of charge decreases; (b) When SOC < 50%, according to P″ ess When SOC>50%, the energy storage battery is charged at P″ ess Multiply by a lambda less than 1 SOC Charging, and SOC It decreases with the increase of state of charge; In charging state, there are In the discharge state, there are Where: SOC is the maximum output constraint coefficient of the energy storage battery, and SOC represents the state of charge of the energy storage battery; The final reference power and state of charge constraint relationship of the energy storage battery is: P ess =P″ ess ×λ SOC Where: P ess It is the final reference power of the energy storage battery.

7. The method for controlling energy storage of an active distribution network containing high-penetration new energy according to claim 1, characterized in that: Analyzing the power response requirements of energy storage participating in grid peak regulation and realizing capacity optimization configuration is to obtain typical daily load curves and daily renewable energy generation curves in different seasons through clustering to form the grid duck curve; selecting typical lines, calculating and analyzing the structure of actual lines, renewable energy installed capacity and user distribution, and establishing a distribution network simulation model; According to the ramp speed and minimum output level of conventional units and the power flow restriction constraints of the tie line, a capacity optimization configuration method for energy storage participating in peak regulation based on particle swarm algorithm is proposed; the peak, valley and flat period division strategy is as follows: calculate the peak, valley and flat membership at each moment, as shown in the following formula: Where: u ft is the duck curve peak membership; u gt is the valley membership of the duck curve; L(t) is the user load at time t; max(L) and min(L) are the load peak and valley values ​​respectively; L c (t) is the output of new energy at time t; max(L c )、min(L c ) are the maximum and minimum values ​​of the new energy output respectively; μ1 and μ2 are the weights of the user load part in the peak and valley membership calculation respectively; In the formula, the greater the peak membership at a certain moment, the greater the possibility that it belongs to the peak period; the greater the valley membership, the greater the possibility that it belongs to the valley period. Therefore, the time period to which the moment belongs is divided by determining the peak valley membership threshold: Where: T f 、T p 、T g They are peak, flat and valley time periods respectively; m1 is the peak membership threshold; m2 is the valley membership threshold.

8. The method for controlling energy storage of an active distribution network containing high-penetration new energy according to claim 1, characterized in that: The optimization dispatch model based on virtual power plant and time-of-use electricity price is divided into two types: The first one: Virtual power plant output optimization model ①Objective function The optimization objective function of the virtual power plant is to minimize the output fluctuation of the virtual power plant: Where: F1 is the volatility index of the virtual power plant output; P VPP.t is the active power of the virtual power plant in period t; P VPP.AV is the average value of the active power of the virtual power plant during the entire dispatch period; in Where: P W (t) is the active power of wind power in period t; P l (t) is the active power of photovoltaic in period t; E ESB (t) is the active power of charging and discharging of the energy storage device; ②Constraints Energy storage device constraints are divided into energy storage charging and discharging power limits and remaining capacity limits; the charging and discharging power limits are as follows: |E ESB (t)|≤E ESB.max Where: E ESB (t) is the charging and discharging active power of the energy storage device in time period t; E ESB.max is the maximum charging and discharging active power of the energy storage device; The remaining capacity is limited to: S min ≤S OC (t)≤S max Where: S oc (t) is the remaining capacity of the energy storage device in period t; S min is the minimum allowable remaining capacity value of the energy storage device; S max The maximum allowable remaining capacity of the energy storage device; Wind power and photovoltaic output constraints Where: P W(t) is the actual active power of the wind farm in period t; P Wf (t) is the predicted active power of the wind farm in period t; P l (t) is the actual active power of the photovoltaic field in period t; P l.f (t) is the predicted active power output of the photovoltaic field in period t; Second. Power grid economic optimization model: In the power grid economic optimization stage, the objective function is to minimize the total cost of system operation. Given the optimized output of the virtual power plant, the power load curve and the operating parameters of each thermal power unit, the optimal output of each thermal power unit is calculated. ①Objective function The economic dispatch objective function of the power grid is: Where: F2 is the operating cost of the entire system; C G.t , C SR.t , C new.t , C ES.t are the thermal power generation cost, spinning reserve cost, renewable energy removal penalty and energy storage device depreciation cost of the system in period t respectively; ②Constraints The climbing power constraint of thermal power unit is: Where: P i (t-1) is the active power of unit i in period t-1; R up i is the maximum rising power of thermal power unit i in a single period; R down i is the maximum power reduction of thermal power unit i in a single period.

9. The method for controlling energy storage of an active distribution network containing high-penetration new energy according to claim 1, characterized in that: The optimization method based on the improved genetic algorithm is: For the first stage of virtual power plant output optimization, the minimum fluctuation is taken as the objective function, and its charging and discharging strategy is determined when the energy storage capacity configuration is determined; For the second stage, the optimal economic dispatching efficiency of the power grid is taken as the objective function. Under the premise that the thermal power units meet the ramp requirements, the maximum output of the new energy units and the energy storage units is taken as the optimization variable for solution.

Images