Virtual energy storage-containing distribution network source-load cooperative region stability control method

Abstract

The invention relates to a virtual energy storage-containing distribution network source-load cooperative region stability control method. The method comprises the following steps of 1, constructing an energy storage system mathematical model, wherein the energy storage system mathematical model comprises a storage battery mathematical model and a virtual energy storage model; 2, constructing a distribution network source load storage collaborative optimization control model containing virtual energy storage; 3, establishing a source load storage system optimization configuration model by adopting a double-layer source load cooperative control strategy, and searching an optimal cooperative control scheme by adopting a particle swarm algorithm to realize source load cooperative optimization scheduling; 4, constructing a source network load storage coordinated optimization evaluation model, and carrying out coordinated scheduling optimization control evaluation through a comprehensive evaluation index; the method has the advantages that a double-layer source-load cooperative control strategy is adopted, the optimal configuration model is constructed, and the virtual energy storage model is established.

Description

Technical Field

[0001] The present invention belongs to the technical field of distribution networks, and in particular relates to a method for regional stability control of source-load coordination in a distribution network containing virtual energy storage. Background Art

[0002] The intermittent nature of renewable energy generation makes energy storage systems an indispensable component for maintaining power balance between the source and load sides in active distribution networks (ADNs). They can effectively alleviate the pressure of system fault recovery when a line fault occurs in the system. However, due to the high cost of traditional energy storage, strict active power regulation of the energy storage system is required to keep it in a safe state of charge (SOC), which increases the safe operation cost of the ADN system. Virtual energy storage technology has been studied to a certain extent both domestically and internationally, mostly based on demand-side load control. In theory, virtual energy storage technology can utilize controllable loads or distributed power sources within the system to reduce the installed capacity and energy storage configuration capacity of traditional power generation equipment, while taking into account the economic efficiency of system operation. However, in the virtual energy storage control technology used in existing technologies, the load power cannot be continuously and smoothly adaptively controlled, and the flexibility is not strong. Therefore, it is very necessary to provide a two-layer source-load coordinated control strategy, construct an optimization configuration model, and establish a virtual energy storage model for a distribution network source-load coordinated regional stability control method containing virtual energy storage. Summary of the Invention

[0003] (1) Technical issues

[0004] In view of the above-mentioned existing technical status, this application mainly addresses the following technical problems:

[0005] 1. Traditional energy storage costs are high, increasing the safe operation costs of the ADN system;

[0006] 2. The virtual energy storage control of the existing technology cannot achieve continuous and smooth adaptive control of load power and has low flexibility.

[0007] (2) Technical solution

[0008] The purpose of the present invention is to overcome the shortcomings of the existing technology and to provide a double-layer source-load collaborative control strategy, construct an optimization configuration model, and establish a virtual energy storage model to achieve a distribution network source-load collaborative regional stability control method containing virtual energy storage.

[0009] The object of the present invention is achieved by: a method for controlling regional stability of a distribution network source-load coordination including virtual energy storage, the method comprising the following steps:

[0010] Step 1: Construct a mathematical model of the energy storage system: including a battery mathematical model and a virtual energy storage model;

[0011] Step 2: Construct a distribution network source-load-storage collaborative optimization control model with virtual energy storage;

[0012] Step 3: Adopt a two-layer source-load collaborative control strategy, establish a source-load-storage system optimization configuration model, and use a particle swarm algorithm to find the optimal collaborative control solution to achieve source-load collaborative optimization scheduling;

[0013] Step 4: Construct a coordinated optimization evaluation model for source, grid, load and storage, and conduct coordinated scheduling optimization control evaluation through comprehensive evaluation indicators.

[0014] Furthermore, the battery mathematical model in step 1 is specifically: the state of charge equation of the charging process is: SOC(t)=(1-δ)SOC(t-1)+P c Δtη c / E bat (1), where δ is the battery self-discharge rate; P c is the charging power; η c is the charging efficiency; E bat is the total amount of energy stored; the state of charge equation during the discharge process is: SOC(t)=(1-δ)SOC(t-1)-P d Δt / (E bat η d )(2), where P d is the discharge power; η d For charging efficiency.

[0015] Furthermore, the virtual energy storage model in step 1 is specifically as follows: when the virtual energy storage control loop is started and not started, the power consumed by the virtual energy storage unit is respectively: Where P′ and P are the power consumed by the pseudo energy storage unit when the control loop is started and not started, respectively; V′ NC 、V NC are the NCL voltages when the control loop is enabled and disabled, respectively; I′ ES , I ES They are the ES output current when the control loop is started and not started respectively; is the NCL impedance angle; therefore, the virtual energy storage power based on ES is: Where ΔP is the virtual energy storage rate; Z NC is the NCL impedance; V S is the bus voltage.

[0016] Furthermore, the virtual energy storage model includes distributed power generation equipment and controllable loads. The distributed power generation equipment includes a wind turbine, which adopts a permanent magnet direct-drive wind turbine generator set and sends wind power to the DC bus side through a voltage source converter; the controllable load includes electric vehicles and seawater desalination loads, among which electric vehicles are connected to the distribution network through a bidirectional converter to realize effective charging and discharging of power batteries.

[0017] Furthermore, the objective function of the optimization control model in step 2 includes two parts. The first objective function is specifically: considering the interests of DG investors and ADN operators, the objective function of the annual operating income of the optimal collaborative scheduling scheme is defined as: maxC = λ1(C sell -C DG )+λ2(C loss +C env -C G -C ES )(6), where C is the annual comprehensive income of DG investors and ADN operators; λ1 and λ2 are the economic index coefficients of DG and ADN respectively; C sell The revenue from electricity sales to DG operators; C DG Investment and operation and maintenance costs for DG operators; C loss ADN loss reduction income; C env ADN environmental benefits; C G The cost of electricity purchased from the main grid by the ADN operator; C ES is the compensation cost of the virtual energy storage unit.

[0018] Furthermore, the second objective function of the optimization control model in step 2 is specifically divided into two parts: benefit and cost, expressed as: Where, F is the total revenue of the distribution network; F1 is the water production revenue of the system; F v is the charging and discharging benefit of electric vehicles; C g It is the cost of air adjustment; C bess are the operation and maintenance costs of wind turbines and hybrid energy storage equipment respectively; C ev It is the peak-shaving cost paid to electric vehicle users.

[0019] Furthermore, in step 3, the particle swarm optimization algorithm is used to solve the first part of the objective function of the optimization control model through a two-level optimization scheduling strategy. Specifically, the upper-level control is implemented by the distributed generation and energy storage system, and the lower-level control is coordinated and optimized based on the dynamic characteristics of the battery and virtual energy storage. The expressions of the particle speed and position of each iteration of the particle swarm optimization algorithm are: Where, and are the velocities of the particle at the kth and k+1th iterations respectively; ω is the inertia weight; c1 and c2 are the weights of the individual extreme value and the global optimal position; r1 and r2 are random numbers; is the individual extreme value of the i-th particle at the k-th iteration; is the global optimal solution at the kth iteration; and are the positions of the particles at the kth and k+1th iterations, respectively. To obtain better local and global optimization capabilities, the inertia weight, individual extreme value, and weight of the global optimal position are set as dynamic parameters: ω = 1 + rand (-0.5, 0.5) (11), c1 = 2 + rand (-1, 1) (12), c2 = 2 + rand (-1, 1) (13), and the improved PS0 algorithm is used to solve the model.

[0020] Furthermore, the particle swarm algorithm in step 3 solves the second part of the objective function of the optimization control model, specifically including the following steps:

[0021] Step A1: Electricity price R = [r1, r2, ..., r 24 ], wind power output power forecast value The charging and discharging power generated by the electric car model Carlo simulation Set the parameters related to the derivation formula and their adjustment ranges, and set the algorithm parameters, including particle size M, maximum number of iterations j, inertia factor w, learning factors c1, c2, and random numbers r1, r2;

[0022] Step A2: Generate the initial population, set the particle position and update speed, assume there are n particles, the position of particle i is the virtual capacitance value of time period t: C v1 (t) i 、C vg (t) i 、C ve (t) i , particle i velocity: v i ;

[0023] Step A3: Calculate the objective function value for each particle. The function with a larger value has higher fitness. Determine the individual extreme value p of the particle. i Group extreme value p g , continuously update the particle position and velocity, and determine the individual extreme value and group extreme value with higher fitness;

[0024] Step A4: If the convergence accuracy is met or the number of iterations is reached, the main loop ends and the output result is the optimal value of the system benefit and the corresponding virtual capacitance value of each time period. Otherwise, return to step A3 and continue iteration.

[0025] Furthermore, the evaluation model in step 4 is specifically as follows: in the comprehensive evaluation process, the evaluation indicators of each link need to be calculated, and then the comprehensive evaluation is achieved by determining the indicator weight of each link. The weight is a quantitative reflection of the importance of the indicator relative to the evaluation target. The comprehensive evaluation indicator value is calculated as follows: Where, F s is the comprehensive evaluation index value; α iis the weight of the bottom i-th indicator relative to the evaluation target; F i is the value of the i-th indicator at the bottom layer; N is the total number of indicators at the bottom layer; β ij is the single weight of the i-th indicator at the bottom layer corresponding to the j-th indicator; n is the total number of layers in the indicator system.

[0026] (3) Beneficial effects

[0027] 1. Introducing virtual energy storage into the energy storage system, a distribution network source-load coordinated regional stability control strategy is proposed: First, a two-layer source-load coordinated control strategy is adopted. The upper layer coordinates distributed generation and energy storage systems, while the lower layer performs coordinated optimization control based on the dynamic characteristics of batteries and virtual energy storage.

[0028] 2. Based on the conversion relationship between the rotational kinetic energy of wind turbines and asynchronous motors and capacitor energy storage, the wind turbine and desalination loads are modeled as virtual energy storage equivalents. Since electric vehicles have both mobile load and energy storage characteristics, an orderly charging and discharging pattern is established using the electricity price mechanism to establish a virtual energy storage model for electric vehicles.

[0029] 3. Combine sources, loads and storage to establish a virtual energy storage system economic optimization model to maximize the daily distribution network operation benefits and form an energy collaborative optimization control strategy with virtual capacitance values ​​and virtual state of charge in each time period as operating parameters; establish a source, load and storage system optimization configuration model and seek the optimal collaborative control solution. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 Schematic diagram of the virtual energy storage structure based on ES of the present invention.

[0031] Figure 2 This is the ES-based virtual energy storage working phasor diagram of the present invention.

[0032] Figure 3 This is the source-load collaborative optimization control flow chart of the present invention.

[0033] Figure 4 This is the electric vehicle load optimization flow chart of the present invention.

[0034] Figure 5 This is a flow chart of the economic optimization calculation of the present invention.

[0035] Figure 6 This is the flow chart of the economic optimization particle swarm algorithm of the present invention.

[0036] Figure 7 This is a flow chart for evaluating the source-grid-load-storage coordination level of the present invention. DETAILED DESCRIPTION

[0037] The present invention will be further described below with reference to the embodiments and / or drawings.

[0038] Example 1

[0039] like Figure 1-7 As shown, a method for regional stability control of source-load coordination in a distribution network including virtual energy storage comprises the following steps:

[0040] Step 1: Construct a mathematical model of the energy storage system: including a battery mathematical model and a virtual energy storage model;

[0041] In the present invention, ① the battery mathematical model: source-load coordinated control requires the energy storage system to perform strict active power regulation to keep it within the safe SOC limit; the state of charge equation during the charging process is: SOC(t)=(1-δ)SOC(t-1)+P c Δtη c / E bat (1), where δ is the battery self-discharge rate; P c is the charging power; η c is the charging efficiency; E bat is the total amount of energy stored; the state of charge equation during the discharge process is: SOC(t)=(1-δ)SOC(t-1)-P d Δt / (E bat η d )(2), where P d is the discharge power; η d For charging efficiency.

[0042] ② Virtual energy storage model: This invention adopts virtual energy storage to participate in ADN source-load coordinated regional control, and introduces a virtual energy storage control method based on electric spring (ES). Its structure is as follows: Figure 1 As shown in the figure, the load with low power quality requirements and can withstand certain voltage fluctuations is non-critical load (NCL), and the load with high power quality requirements is critical load (CL); the voltage source rectifier rectifies the bus voltage to obtain a stable DC voltage; the control loop can obtain the control signal of the inverter according to the upper scheduling requirements. ES-order ;Convert DC voltage to ES output voltage V ES , C d is the DC side capacitance of ES, L f 、C f are the filter inductors and capacitors; taking the non-critical load as a resistive load as an example, the phasor relationship of each parameter in the virtual energy storage equivalent discharge and charging states is as follows Figure 2 As shown, in order to facilitate the analysis of the vector relationship of each parameter, the real part and the imaginary part are added as reference axes, and the specific value is determined according to the specific situation; Figure 2 It can be seen that when the virtual energy storage control loop is started and not started, the power consumed by the virtual energy storage unit is: Where P′ and P are the power consumed by the pseudo energy storage unit when the control loop is started and not started, respectively; V′ NC 、V NC are the NCL voltages when the control loop is enabled and disabled, respectively; I′ ES , I ES They are the ES output current when the control loop is started and not started respectively; is the NCL impedance angle; therefore, the virtual energy storage power based on ES is: Where ΔP is the virtual energy storage rate, ΔP>0 means the virtual energy storage unit is in the charging state, otherwise it is in the discharging state; Z NC is the NCL impedance; V S is the bus voltage.

[0043] Step 2: Construct a distribution network source-load-storage collaborative optimization control model with virtual energy storage;

[0044] In the present invention, the objective function includes two functions: one is to comprehensively consider the interests of DG investors and ADN operators, and define the objective function of the annual operating income of the optimal collaborative scheduling solution as: maxC = λ1(C sell -C DG )+λ2(C loss +C env -C G -C ES )(6), where C is the annual comprehensive income of DG investors and ADN operators; λ1 and λ2 are the economic index coefficients of DG and ADN respectively; C sell The revenue from electricity sales to DG operators; C DG Investment and operation and maintenance costs for DG operators; C loss ADN loss reduction income; C env ADN environmental benefits; C G The cost of electricity purchased from the main grid by the ADN operator; C ES is the compensation cost of the virtual energy storage unit.

[0045] The second part of the objective function is divided into two parts: benefit and cost, expressed as: Where, F is the total revenue of the distribution network; F1 is the water production revenue of the system; F v is the charging and discharging benefit of electric vehicles; C g It is the cost of air adjustment; C bess are the operation and maintenance costs of wind turbines and hybrid energy storage equipment respectively; C ev It is the peak-shaving cost paid to electric vehicle users.

[0046] Load power outage rate: The load power outage rate is inversely proportional to the overall system cost. It is an important indicator for measuring the reliability of the ADN system and is closely related to the economic efficiency of the system. In order to better measure the power supply effect of the system, the load power outage rate is introduced as an evaluation parameter. The expression is: LPSP≤LPSP max (8), where LPSP is the load power shortage rate; LPS(t) is the power difference at time t; E(t) is the load demand at time t; LPSP is the load power shortage rate; LPS(t) is the power difference at time t; E(t) is the load demand at time t; max It is the upper limit of load power failure rate.

[0047] Step 3: Adopt a two-layer source-load collaborative control strategy, establish a source-load-storage system optimization configuration model, and use a particle swarm algorithm to find the optimal collaborative control solution to achieve source-load collaborative optimization scheduling;

[0048] In this invention, ① a two-layer optimization scheduling strategy is adopted to maintain the active power balance and safe and stable operation of the ADN system. The upper-layer control is jointly implemented by the distributed generation and energy storage system. If the battery is repeatedly deeply discharged for a long time, it is easily damaged, resulting in a reduction in cycle life. When the battery energy storage drops to 40% of its capacity, the DG adjustment scheme is activated. By controlling the amount of intermittent DG removal and the output of the controllable DG, the system regulation pressure is effectively relieved, and the deep discharge of the battery is avoided, thereby effectively extending its cycle life.

[0049] The lower-level control is based on the dynamic characteristics of batteries and virtual energy storage for collaborative optimization control. Batteries have high energy density and are convenient for long-term storage of electrical energy. Distributed virtual energy storage units have fast "charging and discharging" speeds, low costs, and large capacity. The collaborative control of the two realizes the distributed coordinated optimization operation of ADN sources and loads.

[0050] ② Improved particle swarm optimization algorithm: Particle swarm optimization algorithm (PS0) has a strong global search capability. Its concept originates from the study of the foraging behavior pattern of bird flocks. The expressions of particle velocity and position in each iteration are: Where, and are the velocities of the particle at the kth and k+1th iterations respectively; ω is the inertia weight; c1 and c2 are the weights of the individual extreme value and the global optimal position; r1 and r2 are random numbers; is the individual extreme value of the i-th particle at the k-th iteration; is the global optimal solution at the kth iteration; and are the positions of the particles at the kth and k+1th iterations respectively.

[0051] In order to obtain better local and global optimization capabilities, the traditional PSO algorithm is improved, and the inertia weight, individual extreme value and global optimal position weight are set as dynamic parameters: ω=1+rand(-0.5,0.5)(11), c1=2+rand(-1,1)(12), c2=2+rand(-1,1)(13). The improved PS0 algorithm is used to solve the model, which has better performance than the traditional algorithm using static parameters. The algorithm process is as follows: Figure 3 As shown, Figure 3 In the example, T is the maximum number of iterations.

[0052] Step 4: Construct a coordinated optimization evaluation model for source, grid, load and storage, and conduct coordinated scheduling optimization control evaluation through comprehensive evaluation indicators.

[0053] In the present invention, ① Evaluation model: During the comprehensive evaluation process, the evaluation indicators of each link need to be calculated, and then the comprehensive evaluation is achieved by determining the indicator weight of each link. The weight is a quantitative reflection of the importance of the indicator relative to the evaluation target. The comprehensive evaluation indicator value is calculated as follows: Where, F s is the comprehensive evaluation index value; α i is the weight of the bottom i-th indicator relative to the evaluation target; F i is the value of the i-th indicator at the bottom layer; N is the total number of indicators at the bottom layer; β ij is the single weight of the i-th indicator at the bottom layer corresponding to the j-th indicator; n is the total number of layers in the indicator system.

[0054] ② Construction of evaluation method: Combining the analytic hierarchy process and the Delphi method, the subjective weight is assigned to the multi-level indicator system, and the direct and indirect factors at different stages from the whole network planning to the dispatching operation are refined to construct a mathematical model of evaluation indicators, such as Figure 7 shown.

[0055] The present invention is a method for regional stability control of source-load coordination of distribution network with virtual energy storage. In use, the present invention proposes a double-layer source-load coordinated control strategy, which collaboratively considers the status of multiple parameters of the ADN system, including the charge state of the energy storage system, the output of photovoltaic and wind turbines, virtual energy storage units, and key load requirements; the upper-layer control adopts distributed power generation and energy storage system coordination, and puts distributed power sources into regulation before the battery charge state reaches the safety lower limit; the lower-layer control performs collaborative control based on the dynamic characteristics of batteries and virtual energy storage; the present invention introduces an ES-based virtual energy storage unit, comprehensively considers the costs of DG operators and ADN operators, establishes an optimization model objective function, and uses an improved PSO algorithm to solve the problem under the premise of ensuring the load power shortage rate. The ADN system with virtual energy storage introduced has reduced energy storage costs, higher comprehensive profits, and a relatively small node voltage offset rate; the present invention has the advantages of a double-layer source-load coordinated control strategy, constructing an optimization configuration model, and establishing a virtual energy storage model.

[0056] Example 2

[0057] like Figure 1-7 As shown, a method for regional stability control of source-load coordination in a distribution network including virtual energy storage comprises the following steps:

[0058] Step 1: Construct a mathematical model of the energy storage system: including a battery mathematical model and a virtual energy storage model;

[0059] In the present invention, the virtual energy storage device includes distributed power generation equipment and controllable loads. The wind turbine adopts a permanent magnet direct-drive wind turbine generator set, and the wind power is sent to the DC bus side through a voltage source converter. The controllable load includes electric vehicles and seawater desalination loads, among which the electric vehicles are connected to the distribution network through a bidirectional converter to realize the effective charging and discharging of the power battery; the seawater desalination load has variable frequency speed regulation capability.

[0060] ① Electric vehicle virtual energy storage model: Electric vehicles have the dual characteristics of load and energy storage. Electric vehicles are connected to the distribution network for orderly charging and discharging. Under the electricity price guidance policy, the charging and discharging of electric vehicles are centrally and orderly controlled on the premise of meeting the usage habits of car owners. Electric vehicles are charged and discharged in a conventional slow manner. Through the Monte Carlo simulation method, by comparing the return time t of the i-th electric vehicle f The end time of the early electricity price trough T ms , the starting time of the evening electricity price peak T ns Before the end of the morning electricity price valley, the electric vehicle virtual energy storage is charged to absorb the redundant power, and after the evening electricity price peak, the electric vehicle virtual energy storage is discharged to reduce the original load power impact; the starting time T of the electric vehicle virtual energy storage charging is reasonably arranged. chars and the starting time of electric vehicle virtual energy storage discharge T dchars for: The discharge time of the electric vehicle's virtual energy storage must ensure that the remaining power can meet the user's daily travel needs, and the discharge amount must not exceed the electric vehicle's maximum discharge depth; the discharge time of the electric vehicle's virtual energy storage T dchar and discharge capacity E dchar Respectively expressed as: Where, C is the total capacity of the electric vehicle power battery; P C is the charging and discharging power of the electric vehicle; w and r are the power consumption per kilometer and the maximum discharge depth of the electric vehicle respectively; S OCEmax 、S OCEmin It is the state of charge limit of electric vehicle battery.

[0061] The charging time of electric vehicles is determined by the discharge amount of electric vehicle virtual energy storage and the power loss during travel. Combined with the constraints on electric vehicles to participate in virtual energy storage at the corresponding electricity price peak and valley times, according to the starting charging and discharging time and discharge duration of electric vehicle virtual energy storage, the end time T of electric vehicle virtual energy storage charging within different return times can be obtained. chare and the end time of electric vehicle virtual energy storage discharge T dchare Expressed as: The above three formulas can be used to calculate the virtual energy storage charge and discharge capacity E of electric vehicles in different periods ev (t) is expressed as: The characteristic curve of electric vehicle charging and discharging demand is guided by electricity price strategy to simulate the load shifting effect of similar energy storage equipment. The orderly charging and discharging scheduling strategy process of electric vehicles is as follows: Figure 3 As shown; the virtual capacitance value C of the electric vehicle at time t can be calculated through the energy relationship ve Expressed as: Where U C is the DC bus terminal voltage.

[0062] The virtual capacitance value can reflect the virtual energy storage capacity of electric vehicles in the regulation. In the storage-load coordinated control, electric vehicles usually maintain the system balance by inputting or removing them only after the energy storage is overcharged or discharged, which may lead to the interruption of dynamic coordination with the energy storage. By adjusting the virtual capacitance value to control the virtual energy storage charge and discharge amount, it is beneficial to formulate the optimal energy storage configuration strategy from the overall level and ensure the advantage of continuous regulation. The virtual state of charge S of electric vehicles OCve Expressed as: Where η v It is the battery charge and discharge efficiency.

[0063] The virtual state of charge is used to evaluate the operating status of virtual energy storage, which is convenient for evaluating the operating status of the system with combined battery energy storage. At the same time, it forms a protection mechanism for the operating status of controllable equipment to prevent the damage caused by transitional regulation to the system. In economic calculations, virtual energy storage usually only uses simplified power indicators and does not form evaluation parameters related to energy storage. The virtual state of charge can establish the same energy storage status evaluation indicators as traditional energy storage. The system benefits obtained by operating under the combined virtual energy storage are not only economically optimal, but also evaluate the operating status of various energy storage by observing the state of charge of the energy storage system, taking into account both economy and safety; the health status of electric vehicles S OHve Expressed as: S OHve (t) = S OHve (t-1)-A×|ΔS OCve |, where A is the linear aging coefficient.

[0064] ② Virtual energy storage model of seawater controllable load: The seawater desalination load is not only suitable for receiving power from distributed power sources, but also the high-pressure pump equipped with a frequency converter can achieve continuous adjustment of the power consumption of the seawater desalination load. It has the typical variable frequency speed regulation performance of asynchronous motors. Taking the asynchronous motor with a relatively large load ratio as an example, the change in rotational kinetic energy corresponding to the change in motor speed has the characteristics of simulating energy storage charging and discharging. The energy conversion relationship between the change in motor speed and the change in supercapacitor voltage is established, which can be expressed as: Where, E r The rotational kinetic energy stored in the rotor; J s 、ω r 、p n are the rotor inertia, electrical angular velocity and pole pair number of the motor respectively; combined with the above formula, the correlation formula of the motor speed is converted into a controllable load virtual energy storage model represented by a variable capacitor; the virtual capacitance value C of the seawater controllable load v1 Expressed as: Where, is the reference value of the electrical angular velocity of the asynchronous motor; ΔU C is the change in the square of the terminal voltage; according to the above formula, the relationship between the angular velocity reference value, the virtual capacitance value and the supercapacitor terminal voltage can be obtained as follows: The virtual energy storage capacitor value of the controllable load changes with the reference value of the electrical angular velocity of the motor rotor, and the electrical angular velocity is affected by the terminal voltage of the supercapacitor; when ΔU C When it increases, As it increases, the motor speed increases to simulate the energy storage charging process; similarly, ΔU C When it decreases, the motor speed decreases to simulate the energy storage discharge process, which to a certain extent shares the regulation pressure of traditional energy storage.

[0065] From the energy perspective, according to the definition of the state of charge of the energy storage element, it can be seen that the virtual state of charge S of the controllable load is:OCv1 for: Where, ω rn is the rated electrical angular velocity of the asynchronous motor; E m is the rotor kinetic energy corresponding to the rated speed.

[0066] ③ Wind turbine virtual energy storage model: Taking a permanent magnet direct-drive wind turbine as an example, the wind turbine's rotational kinetic energy is related to the rotor speed. Since the wind turbine's rotor speed responds quickly, directly controlling the motor's operating state can manage its regulation characteristics. The kinetic energy stored in the rotating mass of the wind turbine can be converted into virtual capacitor charge and discharge energy. The energy conversion relationship can be expressed as: Where, ω g is the electrical angular velocity of the generator; E g Store kinetic energy for the fan rotor; k opt is the maximum power curve coefficient of the unit; the wind turbine operates at the maximum power point to capture the maximum wind energy, and the captured mechanical power is P m ; Combined with the above, the fan virtual capacitance value C vg It is related to the fan speed and bus voltage change, and its expression is: Arranging the above formula, the following formula shows that the change of the generator rotor's rotational kinetic energy has a similar regulation law to that of energy storage. The wind turbine's virtual capacitance value is related to the change of the electrical angular velocity reference value, which can be expressed as: Where, is the reference value of the electrical angular velocity of the wind turbine rotor.

[0067] When the system has excess power, the change in bus voltage is detected, which can be directly reflected in the change in virtual capacitance value through formula (a). According to the speed reference value in formula (b), the wind turbine chooses to slow down. If the power is insufficient and the voltage drops, the wind turbine will operate at the upper limit of the wind speed under maximum power tracking control and cannot further increase the wind power. Therefore, the wind turbine does not start virtual energy storage control at low wind speeds. In order to facilitate MEMS data collection and planning as a whole, the wind turbine virtual state of charge S is introduced. OCvg The energy storage state observation parameters are formed and defined as: Where, ω gn E is the rated electrical angular velocity of the fan rotor; gn is the rotor kinetic energy corresponding to the rated speed.

[0068] Step 2: Construct a distribution network source-load-storage collaborative optimization control model with virtual energy storage;

[0069] In the present invention, the seawater desalination load benefit expression in the second part of the objective function is expressed as: Where k is the price per ton of fresh water; Q is the water production flow rate.

[0070] The adjustment of the wind turbine virtual energy storage charging and discharging power to the original power generates the corresponding wind adjustment cost, which is expressed as: Where K w is the air adjustment coefficient; P wo (t), P w (t) is the wind power output before and after the wind turbine virtual energy storage participates in wind regulation at time t; the wind turbine virtual capacitance value C vg Regulation g (t) The changing trend can realize the control of virtual energy storage power, and the power after wind adjustment P w (t) can be obtained by the following formula:

[0071] Due to the change in air regulation power, the corresponding wind turbine operation and maintenance cost will also change, and its expression is: Where, is the wind turbine operation and maintenance cost per unit electricity.

[0072] By investing in electric vehicle virtual energy storage, arranging electric vehicle users to charge during periods of low electricity prices and discharge during periods of high electricity prices, the operating costs of the distribution network can be reduced. The objective function of the electric vehicle operating income is as follows: Where p t (t) is the electricity price that the distribution network and users interact with during period t; N electric vehicles participate in virtual energy storage scheduling. When charging, users purchase electricity from the microgrid, and when discharging, users sell electricity to the distribution network.

[0073] The participation of electric vehicle virtual energy storage in power regulation will cause battery life loss. To offset part of the battery loss, the distribution network needs to pay users the electric vehicle peak-shaving cost to encourage users to participate in the electric vehicle load-level regulation strategy. The electric vehicle peak-shaving cost expression is: Where, E evd is the discharge amount of electric vehicles participating in virtual energy storage; K ev The electricity price for electric vehicles to be subsidized per unit of discharge.

[0074] The hybrid energy storage device takes the minimum comprehensive operation and maintenance cost as the optimization goal. The comprehensive operation and maintenance cost includes the battery life loss C B Total operation and maintenance costs including supercapacitors and batteries It is expressed as follows: Where, is the operation and maintenance cost per unit of electricity of hybrid energy storage; are the charging and discharging power of the supercapacitor and battery pack during period t respectively.

[0075] Analogous to the virtual energy storage state evaluation method, observe the battery charge state S OCB (t) is expressed as: Where, E B Indicates the remaining capacity of the battery; E Bmaxis the maximum available capacity of the battery; the battery life loss can be determined by the battery state of charge S during period t OCB (t)) indicates; Where K inv is the total investment cost of the battery; S OHmin The health status value of the battery at the end of its life. During the charge and discharge process, the remaining capacity of the battery closely follows the change of its charge and discharge power, and the change pattern is as follows: Where, are the charging and discharging power of the battery pack during period t; η char ,η dis It is the charging and discharging efficiency of the battery pack.

[0076] Constraints: 1) Power balance constraints: 2) Hybrid energy storage constraints: Hybrid energy storage equipment state of charge limit: 20% ≤ S OCB (t)≤80%, 20%≤S OCC (t)≤80%, where S OCC (t) is the state of charge of the supercapacitor during time t; 3) Virtual energy storage related constraints: Electric vehicle battery state of charge and state limit: 30% ≤ S OCve (t)≤90%, 20%≤S OHve (t)≤100%; Asynchronous motor virtual energy storage charge state limit: 21.78%≤S OCv1 (t)≤100%; Wind turbine virtual energy storage charge state limit: 16%≤S OCvg (t)≤100%.

[0077] Step 3: Adopt a two-layer source-load collaborative control strategy, establish a source-load-storage system optimization configuration model, and use a particle swarm algorithm to find the optimal collaborative control solution to achieve source-load collaborative optimization scheduling;

[0078] In the present invention, the particle swarm optimization algorithm (PSO) is used to calculate the optimal economic operation scheme of the virtual energy storage system containing multiple types of controllable resources. The energy management mode process of solving the optimal economic operation scheme is as follows: Figure 5 As shown; the algorithm flow of the PSO algorithm to solve the capacitance value of virtual energy storage at the economically optimal moment t is as follows Figure 6 As shown, the specific steps are:

[0079] Step A1: Electricity price R = [r1, r2, ..., r 24 ], wind power output power forecast value Electric car model Carlo simulates the charging and discharging power generated Set the parameters related to the derivation formula and their adjustment ranges, and set the algorithm parameters, including particle size M, maximum number of iterations j, inertia factor w, learning factors c1, c2, and random numbers r1, r2;

[0080] Step A2: Generate the initial population, set the particle position and update speed, assume there are n particles, the position of particle i is the virtual capacitance value of time period t: C v1 (t) i 、C vg (t) i 、C ve (t) i , particle i velocity: v i ;

[0081] Step A3: Calculate the objective function value for each particle. The function with a larger value has higher fitness. Determine the individual extreme value p of the particle. i Group extreme value p g , continuously update the particle position and velocity, and determine the individual extreme value and group extreme value with higher fitness;

[0082] Step A4: If the convergence accuracy is met or the number of iterations is reached, the main loop ends and the output result is the optimal value of the system benefit and the corresponding virtual capacitance value of each time period. Otherwise, return to step A3 and continue iteration.

[0083] In summary, the Monte Carlo simulation algorithm is used to solve the electric vehicle load model, and electric vehicles realize load transfer in the form of virtual energy storage, reduce the peak-to-valley difference of charging load, and alleviate the impact of electric vehicle charging behavior on the distribution network; the seawater controllable load and the virtual capacitance value of the wind turbine are used as control parameters to realize the conversion of rotor kinetic energy and capacitor energy, reflecting the adjustable characteristics of virtual energy storage; multiple types of controllable sources and loads and energy storage use the same state variables to participate in microgrid regulation, making the system collaborative optimization and monitoring more comprehensive and the collaborative operation integrated; the control strategy proposed in this paper effectively combines MEMS and virtual energy storage technology, and issues collaborative operation instructions represented by capacitor parameters through MEMS. At the same time, the operating status of controllable resources is observed, which can reasonably allocate dynamic scheduling decisions of source, load and storage resources on the basis of maximizing the regulation potential of controllable resources, thereby improving the system operation flexibility and economic potential.

[0084] Step 4: Construct a coordinated optimization evaluation model for source, grid, load and storage, and conduct coordinated scheduling optimization control evaluation through comprehensive evaluation indicators.

[0085] The present invention is a method for regional stable control of source-load coordination of distribution network containing virtual energy storage, and proposes an energy management method for a virtual energy storage system composed of energy storage, controllable load and distributed power generation. In use, a virtual energy storage model is first constructed, and in the intraday optimization stage, a scheduling decision is established with the goal of maximizing the flexibility resource adjustment benefits in the intraday distribution network; secondly, MEMS obtains the optimal energy management strategy and ratio parameters by executing an energy optimization strategy based on virtual energy storage, taking the virtual capacitance value at each moment as a unified control parameter, and taking the virtual charge state of each device as a system state evaluation indicator; the present invention constructs an energy conversion relationship between source-load energy storage and capacitor energy storage for electric vehicles, motors and fans with controllable characteristics in the distribution network, and forms a virtual energy storage collaborative optimization method with virtual capacitance value as an operating parameter and virtual charge state as an evaluation indicator; the present invention has the advantages of a double-layer source-load collaborative control strategy, constructing an optimization configuration model, and establishing a virtual energy storage model.

Claims

1. A method for regional stability control of source-load coordination in a distribution network with virtual energy storage, characterized by: The method comprises the following steps: Step 1: Construct a mathematical model of the energy storage system: including a battery mathematical model and a virtual energy storage model; Step 2: Construct a distribution network source-load-storage collaborative optimization control model with virtual energy storage; Step 3: Adopt a two-layer source-load collaborative control strategy, establish a source-load-storage system optimization configuration model, and use a particle swarm algorithm to find the optimal collaborative control solution to achieve source-load collaborative optimization scheduling; Step 4: Construct a coordinated optimization evaluation model for source, grid, load and storage, and conduct coordinated scheduling optimization control evaluation through comprehensive evaluation indicators.

2. The method for regional stability control of a distribution network source-load coordination system with virtual energy storage according to claim 1, characterized in that: The battery mathematical model in step 1 is specifically: the state of charge equation during the charging process is: SOC(t)=(1-δ)SOC(t-1)+P c Δtη c / E bat (1), where δ is the battery self-discharge rate; P c is the charging power; η c is the charging efficiency; E bat is the total amount of energy storage; The state of charge equation during the discharge process is: SOC(t)=(1-δ)SOC(t-1)-P d Δt / (E bat η d )(2), where P d is the discharge power; η d For charging efficiency.

3. The method for regional stability control of a distribution network source-load coordination system with virtual energy storage according to claim 2, characterized in that: The virtual energy storage model in step 1 is specifically as follows: when the virtual energy storage control loop is activated and not activated, the power consumed by the virtual energy storage unit is respectively: Where P′ and P are the power consumed by the pseudo energy storage unit when the control loop is started and not started, respectively; V′ NC 、V NC are the NCL voltages when the control loop is enabled and disabled, respectively; I′ ES , I ES They are the ES output current when the control loop is started and not started respectively; is the NCL impedance angle; therefore, the virtual energy storage power based on ES is: Where ΔP is the virtual energy storage rate; Z NC is the NCL impedance; V S is the bus voltage.

4. The method for regional stability control of a distribution network source-load coordination system with virtual energy storage according to claim 3, characterized in that: The virtual energy storage model includes distributed power generation equipment and controllable loads. The distributed power generation equipment includes a wind turbine, which adopts a permanent magnet direct-drive wind turbine generator set and transmits wind power to the DC bus side through a voltage source converter. The controllable load includes electric vehicles and seawater desalination loads, among which electric vehicles are connected to the microgrid through a bidirectional converter to realize the effective charging and discharging of power batteries.

5. The method for regional stability control of distribution network source-load coordination with virtual energy storage according to claim 1, characterized in that: The objective function of the optimization control model in step 2 includes two parts. The first objective function is specifically: considering the interests of DG investors and ADN operators, the objective function of the annual operating income of the optimal collaborative scheduling solution is defined as: maxC = λ1(C sell -C DG )+λ2(C loss +C env -C G -C ES )(6), where C is the annual comprehensive income of DG investors and ADN operators; λ1 and λ2 are the economic index coefficients of DG and ADN respectively; C sell The income from electricity sales for DG operators; C DG Investment and operation and maintenance costs for DG operators; C loss ADN loss reduction income; C env ADN environmental benefits; C G The cost of electricity purchased from the main grid by the ADN operator; C ES is the compensation cost of the virtual energy storage unit.

6. The method for regional stability control of a distribution network source-load coordination system with virtual energy storage according to claim 5, characterized in that: The second objective function of the optimization control model in step 2 is specifically divided into two parts: benefit and cost, expressed as: Where, F is the total revenue of the distribution network; F1 is the water production revenue of the system; F v is the charging and discharging benefit of electric vehicles; C g It is the cost of air adjustment; C bess are the operation and maintenance costs of wind turbines and hybrid energy storage equipment respectively; C ev It is the peak-shaving cost paid to electric vehicle users.

7. The method for regional stability control of a distribution network source-load coordination system with virtual energy storage according to claim 5, characterized in that: In step 3, the particle swarm optimization algorithm is used to solve the first part of the objective function of the optimization control model through a two-level optimization scheduling strategy. Specifically, the upper-level control is implemented by the distributed generation and energy storage system, and the lower-level control is coordinated and optimized based on the dynamic characteristics of the battery and virtual energy storage. The expressions of the particle speed and position of each iteration of the particle swarm optimization algorithm are: Where, and are the velocities of the particle at the kth and k+1th iterations respectively; ω is the inertia weight; c1 and c2 are the weights of the individual extreme value and the global optimal position; r1 and r2 are random numbers; is the individual extreme value of the i-th particle at the k-th iteration; is the global optimal solution at the kth iteration; and are the positions of the particles at the kth and k+1th iterations, respectively. To obtain better local and global optimization capabilities, the inertia weight, individual extreme value, and weight of the global optimal position are set as dynamic parameters: ω = 1 + rand (-0.5, 0.5) (11), c1 = 2 + rand (-1, 1) (12), c2 = 2 + rand (-1, 1) (13), and the improved PS0 algorithm is used to solve the model.

8. The method for regional stability control of distribution network source-load coordination with virtual energy storage according to claim 6, characterized in that: The particle swarm algorithm in step 3 solves the second part of the objective function of the optimization control model, specifically including the following steps: Step A1: Electricity price R = [r1, r2, ..., r 24 ], wind power output power forecast value Electric car model Carlo simulates the charging and discharging power generated Set the parameters related to the derivation formula and their adjustment ranges, and set the algorithm parameters, including particle size M, maximum number of iterations j, inertia factor w, learning factors c1, c2, and random numbers r1, r2; Step A2: Generate the initial population, set the particle position and update speed, assume there are n particles, the position of particle i is the virtual capacitance value of time period t: C v1 (t) i 、C vg (t) i 、C ve (t) i , particle i velocity: v i ; Step A3: Calculate the objective function value for each particle. The function with a larger value has higher fitness. Determine the individual extreme value p of the particle. i Group extreme value p g , continuously update the particle position and velocity, and determine the individual extreme value and group extreme value with higher fitness; Step A4: If the convergence accuracy is met or the number of iterations is reached, the main loop ends and the output result is the optimal value of the system benefit and the corresponding virtual capacitance value of each time period. Otherwise, return to step A3 and continue iteration.

9. The method for regional stability control of distribution network source-load coordination with virtual energy storage according to claim 1, characterized in that: The evaluation model in step 4 is specifically as follows: in the comprehensive evaluation process, the evaluation indicators of each link need to be calculated, and then the comprehensive evaluation is achieved by determining the indicator weight of each link. The weight is a quantitative reflection of the importance of the indicator relative to the evaluation target. The comprehensive evaluation indicator value is calculated as follows: Where, F s is the comprehensive evaluation index value; α i is the weight of the bottom i-th indicator relative to the evaluation target; F i is the value of the i-th indicator at the bottom layer; N is the total number of indicators at the bottom level; β ij is the single weight of the i-th indicator at the bottom layer corresponding to the j-th indicator; n is the total number of layers in the indicator system.

Images