Active power distribution network all-vanadium redox flow battery energy storage system nested double-layer optimization operation method considering dynamic performance

Abstract

The invention discloses an active power distribution network all-vanadium redox flow battery energy storage system nested double-layer optimization operation method considering dynamic performance. The method comprises the following steps: establishing a mathematical model of charge-discharge efficiency, dynamic self-discharge rate and dynamic capacity fading performance of an all-vanadium redox flow battery, and on the basis, constructing a nested double-layer optimization operation model of the active power distribution network battery energy storage system considering the dynamic performance by adopting a meta-model heuristic method, wherein the inner-layer optimization target is the dynamic performance of the all-vanadium redox flow battery, the outer-layer optimization target is the comprehensive operation income of the active power distribution network, an improved particle swarm optimization algorithm is adopted to solve a nested double-layer optimization operation scheme of the energy storage system, and an energy storage optimization operation scheme is constructed under the collaborative optimization of the battery performance and economy. The invention provides a nested double-layer optimization operation model of an all-vanadium redox flow battery energy storage system of an active power distribution network considering dynamic performance, solves the problem of collaborative optimization of the dynamic performance of the battery and the comprehensive operation benefit of the active power distribution network, and provides a reasonable operation scheme of the battery energy storage system of the active power distribution network.

Description

Technical Field

[0001] The present invention belongs to the technical field of electric power engineering, and is specifically a nested double-layer optimization operation method for an active distribution network all-vanadium liquid flow battery energy storage system taking into account dynamic performance. Background Art

[0002] Renewable energy has become the most widely used secondary energy source due to its advantages such as economical transportation and no pollution. Energy storage system has attracted much attention as a bottleneck technology with a guiding role in the field of power system. Renewable energy such as solar energy and wind energy are greatly affected by climate and environmental factors, and direct grid connection will bring many problems. Summary of the invention

[0003] In view of the deficiencies in the prior art, the present invention proposes a nested double-layer optimization operation method for an active distribution network all-vanadium liquid flow battery energy storage system taking into account dynamic performance.

[0004] In order to achieve the above-mentioned purpose, the present invention adopts the following technical scheme: a nested double-layer optimization operation method of an active distribution network all-vanadium liquid flow battery energy storage system taking into account dynamic performance, comprising the following steps:

[0005] Step 1: Model the dynamic performance of the all-vanadium liquid flow battery by performing a charge and discharge test on the all-vanadium liquid flow battery;

[0006] Step 2: Model the dynamic output characteristics of the energy storage battery through the energy and power changes of the all-vanadium liquid flow battery;

[0007] Step 3: Establish an active distribution network comprehensive operation benefit indicator system by analyzing the functional and economic effects of energy storage systems connected to the active distribution network from multiple perspectives;

[0008] Step 4: construct a nested two-layer optimization operation model;

[0009] Step 5: Use the optimized particle swarm algorithm to solve the nested double-layer optimization operation model to obtain the nested double-layer optimization operation plan of the energy storage system.

[0010] Preferably, the specific method for modeling the dynamic performance of the all-vanadium liquid flow battery in step 1 is:

[0011] Through the charging and discharging power experimental test of the all-vanadium liquid flow battery, the battery efficiency data matrix set A is constructed, and the charging efficiency η of the all-vanadium liquid flow battery is calculated. c and discharge efficiency η d The least squares method is used for fitting to obtain the battery charge and discharge efficiency model, which are as follows:

[0012]

[0013] Where Tb is the operating temperature, P b is the operating efficiency, SoC is the state of charge of the energy storage system, a c , b c 、c c ,d c and a o 、c o 、a c1 , b c1 、c c1 ,d c1 is the VRB charging efficiency coefficient fitting parameter value, a d , b d 、c d with a d1 、a o 、c d1 , b o ,d o is the VRB discharge efficiency coefficient fitting parameter value;

[0014] Step 1.2: Through the static experimental test of the all-vanadium liquid flow battery under different SoCs, the battery self-discharge rate data matrix set B is constructed, and the battery self-discharge rate γ is fitted to obtain the battery self-discharge rate model, which is specifically:

[0015] γ=(m a T b +m b )SoC·(m c SoC+m d )

[0016] Among them, T b is the operating temperature, m a 、m b 、m c 、m d All are fitting parameter values ​​of self-discharge rate coefficient;

[0017] Step 1.3, obtain the semi-empirical formula of battery capacity decay performance, fit the relevant parameters through the battery charge and discharge capacity decay experimental test, and obtain the battery dynamic capacity decay performance model, specifically:

[0018]

[0019] In the formula, k Nc is the cycle number coefficient, Nc n is the number of cycles equivalent to 100% charge and discharge cycle, N is the number of experiments, and the average value of the state of charge in a single charge and discharge cycle is defined as SoC a , using the mean square error of the state of charge SoC d Describes the fluctuation amplitude of the state of charge during the cycle, τ is the time constant, where kT are the temperature influence coefficients, T and T ref are the actual value and reference value of the ambient temperature, T a,t and T refa are the actual value and reference value of the internal operating temperature of VRB, respectively; k1, k2, k3, and k4 are the fitting coefficients of the relationship between capacity attenuation and state of charge.

[0020] Preferably, step 2 models the dynamic output characteristics of the energy storage battery, and the model is specifically:

[0021]

[0022] Where, SoC t and SoC t-1 They represent the state of charge at time t and time t-1 respectively, E VRB represents the capacity of VRB, γ is the self-discharge rate, α is the capacity decay rate, η c / d represents the charging or discharging efficiency of the all-vanadium liquid flow battery, P VRB Output power of the all-vanadium liquid flow battery energy storage system.

[0023] Preferably, the specific process of establishing the comprehensive operation benefit indicator system of the active distribution network is as follows:

[0024] Step 3.1: Consider the abandoned PV as the operating income of the active distribution network and determine the PV consumption:

[0025] Photovoltaic consumption B ab The calculation is as follows:

[0026] B ab =k a W VRB

[0027] In the formula, k a is the price factor, W VRB is the amount of abandoned photovoltaic power generation, where the total amount of abandoned photovoltaic power generation is calculated as follows:

[0028]

[0029] Where P wind is the power generation of the wind turbine in the energy storage system, N wind is the installed capacity of wind power, P solar is the photovoltaic power generation in the energy storage system, N solar is the number of photovoltaic installed capacity, P grid_min P is the minimum value of the amount of electricity that the energy storage system purchases from the upper power grid when the active distribution network is running. loss is the network loss, P load is the load power, P VRB is the output power of the energy storage system, Zt It is the characteristic quantity of photovoltaic abandonment state of energy storage system;

[0030] Step 3.2: Peak shaving and valley filling is the operating benefit of the active distribution network. Calculate the peak shaving and valley filling benefit B up The specific method is as follows:

[0031] Calculate the peak shaving rate ω, the specific formula is:

[0032]

[0033] Where P max is the load peak value during the energy storage operation before peak shaving, P' max It is the load peak value during the energy storage operation after peak shaving;

[0034] Calculate the number of years N for delaying grid transformation and upgrading based on the peak reduction rate y , the specific calculation formula is as follows:

[0035]

[0036] Where, ε is the annual load growth rate;

[0037] The benefits of delaying the transformation and upgrading obtained by optimizing the energy storage operation are calculated based on the number of years of delaying the transformation and upgrading of the power grid. The formula is as follows:

[0038]

[0039] In the formula, C e The investment required for grid transformation and upgrading is r With d r are the interest rate and discount rate respectively;

[0040] Step 3.3: Calculate the power market arbitrage B arb :

[0041]

[0042] In the formula, τ is the total year, Δf is the difference in the operating income of the distribution network in a year, and C p is the real-time electricity price, P sell is the electricity sold by the energy storage system. When the energy storage system sells electricity to the load, P sell >0, when purchasing electricity P sell <0;P grid The energy storage system reduces the amount of electricity purchased from the upper grid by storing photovoltaic power;

[0043] Step 3.4: Calculate environmental benefits B em :

[0044]

[0045] In the above formula, C w,k Charge fees to each unit that emits pollutants; grid,k is the pollutant emission density of the upper power grid; P g It represents the total amount of electricity purchased from the upper power grid that is reduced when the energy storage system is running, M represents the type of pollutant, and P grid It indicates the total amount of electricity purchased from the upper power grid that is reduced every hour when the energy storage system is running.

[0046] Preferably, a metamodel heuristic method is adopted, with the dynamic performance of the all-vanadium liquid flow battery as the inner optimization target and the comprehensive operation benefit of the active distribution network as the outer optimization target, to construct a nested double-layer optimization operation model, wherein the inner and outer objective functions of the nested double-layer optimization operation model are:

[0047] Inner objective function F1 = max{ω1η c / d -ω2λ-ω3α};

[0048] Outer objective function F2 = max{β1B ab +β2B up +β3B arb +β4B em}

[0049] Among them, ω1, ω2, ω3, β1, β2, β3, and β4 are the weight values ​​of the corresponding decision variables.

[0050] Preferably, the optimized particle swarm algorithm is used to solve the nested double-layer optimization operation model, and the specific process of obtaining the nested double-layer optimization operation scheme of the energy storage system is as follows:

[0051] Step 5.1: Add constraints to the nested two-layer optimization model, including:

[0052] For a small distribution network containing photovoltaic power generation and wind power generation, the power balance satisfies the following equation:

[0053]

[0054] Among them, P grid is the power transmitted from the upper power grid to this area, P wind The power input to the region for wind power generation in the region, P solar The power input to the area for photovoltaic power generation, P VRB is the power of the energy storage device;

[0055] 2) Due to the capacity limitation of the transformer, the power P input into this area from the upper level power grid is grid Will be restricted to a certain range, namely:

[0056] P grid_min ≤P grid ≤P grid_max -

[0057] Among them, P grid_min Enter the minimum power value of the area for the upper level power grid, P grid_max Input the maximum power of the area to the upper level power grid;

[0058] 3) SoC for energy storage battery systems t Between maximum state of charge SoC max and minimum state of charge SoC min Between

[0059] SoC min ≤SoC t ≤SoC max

[0060] In the energy storage system, SoC is used to reflect the remaining capacity of the energy storage battery. The calculation method is as follows:

[0061]

[0062] Among them, E VRB is the rated capacity of the energy storage system battery, P VRB is the real-time power of the energy storage system. VRB When P is greater than 0, the energy storage system discharges. VRB When it is less than 0, the energy storage battery is charged;

[0063] Step 5.2: Use the improved particle swarm algorithm to solve the nested two-layer optimization operation model. The specific process is as follows:

[0064] (1) Collect data, including photovoltaic power generation, wind power generation, and power load;

[0065] (2) Initialize the particle swarm, take the real-time power and charge state of each energy storage as a particle, and assign it an initial position and initial velocity;

[0066] (3) According to the charge and discharge efficiency, capacity attenuation rate, self-discharge rate and economic benefit index function that need to be solved, the fitness of each particle is calculated by the inner layer F1 and the outer layer objective function F2 of the nested double-layer optimization operation model; according to the fitness value of each particle, the global optimal position is solved;

[0067] (4) Update the particle position and particle velocity;

[0068] (5) Determine whether the number of iterations is met. If not, continue to update the particle fitness and position. If so, the calculation ends and the best position, i.e., the optimal solution, is obtained.

[0069] Compared with the prior art, the present invention has the following significant advantages: the present invention combines the dynamic performance of the all-vanadium liquid flow battery with the optimal operating benefit of the active distribution network, and adopts an improved particle swarm algorithm, which greatly increases the operating benefit of the all-vanadium liquid flow battery energy storage system of the active distribution network. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] Figure 1 This is the dynamic efficiency curve of all-vanadium liquid flow battery.

[0071] Figure 2 This is the dynamic self-discharge rate curve of all-vanadium liquid flow battery.

[0072] Figure 3 This is the dynamic capacity decay performance curve of all-vanadium liquid flow battery.

[0073] Figure 4 Flowchart for modeling an all-vanadium flow battery energy storage system.

[0074] Figure 5 Modeling the model solution process for an all-vanadium liquid flow battery energy storage system.

[0075] Figure 6 It is a line graph of typical daily load demand and electricity price curve.

[0076] Figure 7 It is a line chart of daily photovoltaic output, wind turbine output and renewable energy output. DETAILED DESCRIPTION

[0077] A nested double-layer optimization operation method for an active distribution network all-vanadium liquid flow battery energy storage system taking into account dynamic performance, the specific steps are:

[0078] Step 1: Through the charging and discharging test of the all-vanadium liquid flow battery, the dynamic performance of the all-vanadium liquid flow battery is modeled, which is specifically described as follows:

[0079] Step 1.1: Through the charging and discharging power experimental test of the all-vanadium liquid flow battery, the equivalent loss analysis of the all-vanadium liquid flow battery is carried out, the heat loss power and output power of the battery at different charging and discharging rates are analyzed, and the battery efficiency data matrix set A is constructed. Then, the charging efficiency η of the all-vanadium liquid flow battery is calculated according to formula (1) and formula (2). c and discharge efficiency η d The least squares method is used for fitting to obtain the relevant parameter values ​​of the battery dynamic efficiency model.

[0080]

[0081] Where T b is the operating temperature, P b is the operating efficiency, SoC is the state of charge of the energy storage system, a c 、b c 、c c d c and a o 、c o 、a c1 、b c1 、c c1 d c1 is the VRB charging efficiency coefficient fitting parameter value, a d 、b d 、c d with a d1 、a o 、c d1 、b o d o is the VRB discharge efficiency coefficient fitting parameter value. The specific parameter values ​​are shown in Table 1.

[0082] Table 1. Relevant parameter values ​​of battery dynamic efficiency model

[0083] a b C d C -0.1280 1.0500 0.0380 -0.1180 c1 9.4750 7.7600 6.8200 46.7900 d 1.0334 0.3454 0.1192 d1 -7.6000 6.3606 52.3350 o 0.0380 1.1755 63.2674 51.2653

[0084] Step 1.2: Through static experimental tests of all-vanadium liquid flow batteries under different SoCs, the battery self-discharge rate data matrix set B is constructed according to the time and rate analysis of the battery self-discharge to the terminal, and then the correlation coefficient of the battery self-discharge rate γ is fitted according to formula (3) to obtain the fitting coefficient.

[0085] γ=(m a T b +m b )SoC·(m c SoC+m d ) (3)

[0086] Where, Tb is the operating temperature, m a 、m b 、m c 、m d All are fitting parameter values ​​of self-discharge rate coefficient. Specific parameter values ​​are shown in Table 2.

[0087] Table 2. Related parameter values ​​of battery dynamic self-discharge rate model

[0088] <![CDATA[m a ]]> <![CDATA[m b ]]> <![CDATA[m c ]]> <![CDATA[m d ]]> 0.0211 -0.6099 -0.6751 0.6655

[0089] Step 1.3: Test the battery charge and discharge capacity attenuation experiment, taking into account the temperature, state of charge and number of cycles N. cRegarding the impact of battery capacity decay, the capacity decay rate of the all-vanadium liquid flow battery is fitted according to the semi-empirical formula, and formula (4) is obtained:

[0090]

[0091] In the formula, k Nc is the cycle number coefficient, Nc n is the number of cycles equivalent to 100% charge and discharge cycle, N is the number of experiments, and the average value of the state of charge in a single charge and discharge cycle is defined as SoC a , using the mean square error of the state of charge SoC d Describes the fluctuation amplitude of the state of charge during the cycle, τ is the time constant, where k T are the temperature influence coefficients, T and T ref are the actual value and reference value of the ambient temperature, T a,t and T refa are the actual value and reference value of the internal operating temperature of VRB, k1, k2, k3, k4 are the fitting coefficients of the relationship between capacity attenuation and state of charge, and the specific parameter values ​​are shown in Table 3.

[0092] Table 3. Relevant parameter values ​​of capacity decay model of all-vanadium liquid flow battery

[0093] <![CDATA[k1]]> <![CDATA[k2]]> <![CDATA[k3]]> -4.09e-02 -2.167 1.41e-03 <![CDATA[k4]]> <![CDATA[k T1 ]]> <![CDATA[k T2 ]]> 6.13 0.0254 0.0937 <![CDATA[T ref ]]> <![CDATA[T refa ]]> <![CDATA[k Nc ]]> 25 425 3.17e-4

[0094] Step 2: Considering the influence of battery efficiency, capacity attenuation, and self-discharge rate on battery SoC, the dynamic output characteristic model of the energy storage battery is obtained according to formula (5):

[0095]

[0096] Where, SoC t and SoC t-1 They represent the state of charge at time t and time t-1 respectively, E VRB represents the capacity of VRB, γ is the self-discharge rate, α is the capacity decay rate, η c / d represents the charging or discharging efficiency of the all-vanadium liquid flow battery, P VRB Output power of the all-vanadium liquid flow battery energy storage system.

[0097] Step 3: Establish an active distribution network comprehensive operation benefit indicator system. The specific steps are as follows:

[0098] Step 3.1: The total abandoned PV power reflects the level of PV consumption. The energy storage system can store electricity during the period of excess PV. Consider the abandoned PV as the operating income of the active distribution network, where PV consumption B ab According to formulas (6) and (7), the equivalent discount is calculated by the total abandoned photovoltaic capacity and price factor.

[0099] The total abandoned PV capacity can be calculated as follows:

[0100]

[0101] In formula (6) and formula (7), P wind is the power generation of the wind turbine in the energy storage system, N wind is the installed capacity of wind power, P solar is the photovoltaic power generation in the energy storage system, N solar is the number of photovoltaic installed capacity, P grid_min P is the minimum value of the amount of electricity that the energy storage system purchases from the upper power grid when the active distribution network is running. loss is the network loss, P load is the load power, P VRB is the output power of the energy storage system, Z t It is the characteristic quantity of photovoltaic abandonment state of energy storage system.

[0102] Photovoltaic consumption B ab The calculation is as follows:

[0103] B ab =k a W VRB (8)

[0104] In the formula, k a is the price factor, W VRB For abandoned photovoltaic power.

[0105] Step 3.2: Consider peak shaving and valley filling as the operating benefit of the active distribution network. By charging at low load and discharging at peak load, the energy storage system can reduce the peak-to-valley difference of load, smooth the load curve, reduce the pressure on the grid, and thus delay the transformation and upgrading of the grid, which indirectly brings economic benefits. Peak shaving and valley filling benefit B up According to formula (9), the equivalent discount income is:

[0106] The peak clipping rate ω can be calculated using the following formula:

[0107]

[0108] In the above formula, P max and P' max It is the difference of load peak value during the energy storage operation.

[0109] The number of years to delay grid transformation and upgrading is N y , the specific calculation formula is as follows:

[0110]

[0111] Where ε is the annual load growth rate.

[0112] The calculation formula for the benefits of delaying transformation and upgrading obtained by optimizing energy storage operation is as follows:

[0113]

[0114] In the formula, C e The investment required for grid transformation and upgrading is r With d r The interest rate and discount rate

[0115] Step 3.3: When the photovoltaic power generation is excessive or the electricity price is low, the energy storage system purchases electricity from the upper grid and sells electricity when the load demand of the active distribution network is large or the electricity price is high according to the operation strategy. Therefore, the arbitrage income in the electricity market is B arb It can be calculated by equation (12) and equation (13):

[0116]

[0117] In the above formula, τ is the total year, Δf is the difference in the operating income of the distribution network in a year, and C p is the real-time electricity price, P sell is the electricity sold by the energy storage system. When the energy storage system sells electricity to the load, P sell >0, when purchasing electricity P sell <0;P grid The energy storage system reduces the amount of electricity purchased from the upper grid by storing photovoltaic power.

[0118] Step 3.4, Environmental Benefit B em The equivalent discount income is as follows:

[0119] In areas with high photovoltaic penetration, energy storage systems can charge during periods of high light intensity and low electricity consumption, and discharge during periods of peak electricity consumption, thereby reducing the amount of electricity purchased by the power grid at this level from the upper power grid and achieving greenhouse gas emission reduction benefits. The environmental benefit expression of energy storage is as follows:

[0120]

[0121] In the above formula, C w,k Charge fees to each unit that emits pollutants; grid,k is the pollutant emission density of the upper power grid; P g It represents the total amount of electricity purchased from the upper power grid that is reduced when the energy storage system is running, M represents the type of pollutant, and P grid It indicates the total amount of electricity purchased from the upper power grid that is reduced every hour when the energy storage system is running.

[0122] Step 4: Using the metamodel heuristic method, the dynamic performance of the all-vanadium liquid flow battery is taken as the inner optimization target, and the comprehensive operation benefit of the active distribution network is taken as the outer optimization target to construct a nested two-layer objective function, where the inner and outer objective functions are:

[0123] Inner objective function F1 = max{ω1η c / d -ω2λ-ω3α}; (16) Outer objective function F2 = max{β1B ab +β2B up +β3B arb +β4B em} (17)

[0124] Among them, ω1, ω2, ω3, β1, β2, β3, and β4 are the weight values ​​of the corresponding decision variables, and the weight of each indicator is calculated using the analytic hierarchy process (AHP) evaluation method.

[0125] Step 5: Use the optimized particle swarm algorithm to solve the nested double-layer optimization operation model and obtain the nested double-layer optimization operation plan for the energy storage system. The solution process is as follows: Figure 5 shown.

[0126] Step 5.1: In order to solve the minimum dynamic loss and maximum economic benefit of the energy storage system, it is necessary to add constraints to the nested two-layer optimization operation model.

[0127] 1) For a small distribution network containing photovoltaic power generation and wind power generation, the power balance satisfies the following equation:

[0128]

[0129] Among them, P grid is the power transmitted from the upper power grid to this area, P wind The power input to the region for wind power generation in the region, P solar The power input to the area for photovoltaic power generation, P VRB is the power of the energy storage device.

[0130] 2) Due to the capacity limitation of the transformer, the power P input into this area from the upper level power grid is grid will be restricted to a certain range.

[0131] P grid_min ≤P grid ≤P grid_max (19)

[0132] Among them, P grid_min Enter the minimum power value of the area for the upper level power grid, P grid_max Enter the maximum power value for this area for the upper grid.

[0133] 3) SoC for energy storage battery systems t (State of Charge) should be between the maximum state of charge and the minimum state of charge.

[0134] SoC min ≤SoC t ≤SoC max (20)

[0135] Among them, SoC min SoC is the minimum state of charge of the energy storage battery system. max It is the maximum state of charge of the energy storage battery system.

[0136] In the energy storage system, SoC is used to reflect the remaining capacity of the energy storage battery. The calculation method is as follows:

[0137]

[0138] Among them, E VRB is the rated capacity of the energy storage system battery, P VRB is the real-time power of the energy storage system. VRB When P is greater than 0, the energy storage system discharges. VRB When it is less than 0, the energy storage battery is charged.

[0139] Step 5.2: Use the improved particle swarm algorithm to solve and analyze the two-layer optimization model. The specific process is as follows:

[0140] (1) Collect data, such as photovoltaic power generation, wind power generation, power load, etc.

[0141] (2) Initialize the particle swarm, take the real-time power and charge state of each energy storage as a particle, and assign it an initial position and initial velocity;

[0142] (3) According to the charge and discharge efficiency, capacity attenuation rate, self-discharge rate and economic benefit index function that need to be solved, the economic benefit index function is first fixed, the fitness of the inner layer F1 of the double-layer optimization operation model is calculated, the particle position and particle speed are updated, and the adaptive inertia weight parameter is updated. In the next stage, the economic benefit index is updated by the particle position solved cyclically in the previous stage, and then the fitness of the outer layer F2 of the double-layer optimization operation model is calculated.

[0143] (6) Update the particle position and particle velocity.

[0144] (7) Determine whether the global optimal solution meets the judgment conditions, and at the same time check whether all constraints are met or whether the number of iterations is met. If not, continue to update the particle fitness and position. If so, the calculation ends and the best position, i.e., the optimal solution, is obtained.

[0145] This patent aims at the actual working conditions and operation requirements of regional energy storage, and provides a nested double-layer optimization operation method for the active distribution network all-vanadium liquid flow battery energy storage system taking into account the dynamic performance. It constructs an energy storage optimization operation plan under the coordinated optimization of battery performance and economy, and achieves the coordinated optimization of battery dynamic performance and the comprehensive operation benefits of the active distribution network.

Claims

1. A nested double-layer optimization operation method for an active distribution network all-vanadium liquid flow battery energy storage system taking into account dynamic performance, characterized in that: The steps include: Step 1: Model the dynamic performance of the all-vanadium liquid flow battery by performing a charge and discharge test on the all-vanadium liquid flow battery; Step 2: Model the dynamic output characteristics of the energy storage battery through the energy and power changes of the all-vanadium liquid flow battery; Step 3: Establish an active distribution network comprehensive operation benefit indicator system by analyzing the functional and economic effects of energy storage systems connected to the active distribution network from multiple perspectives; Step 4: construct a nested two-layer optimization operation model; Step 5: Use the optimized particle swarm algorithm to solve the nested double-layer optimization operation model to obtain the nested double-layer optimization operation plan of the energy storage system.

2. The nested double-layer optimization operation method of the active distribution network all-vanadium liquid flow battery energy storage system taking into account dynamic performance according to claim 1 is characterized in that: The specific method for modeling the dynamic performance of the all-vanadium liquid flow battery in step 1 is: Through the charging and discharging power experimental test of the all-vanadium liquid flow battery, the battery efficiency data matrix set A is constructed, and the charging efficiency η of the all-vanadium liquid flow battery is calculated. c and discharge efficiency η d The least squares method is used for fitting to obtain the battery charge and discharge efficiency model, which are as follows: Where T b is the operating temperature, P b is the operating efficiency, SoC is the state of charge of the energy storage system, a c , b c 、c c ,d c and a o 、c o 、a c1 , b c1 、c c1 ,d c1 is the VRB charging efficiency coefficient fitting parameter value, a d , b d 、c d with a d1 、a o 、c d1 , b o ,d o is the VRB discharge efficiency coefficient fitting parameter value; Step 1.2: Through the static experimental test of the all-vanadium liquid flow battery under different SoCs, the battery self-discharge rate data matrix set B is constructed, and the battery self-discharge rate γ is fitted to obtain the battery self-discharge rate model, which is specifically: γ=(m a T b +m b )SoC·(m c SoC+m d ) Among them, T b is the operating temperature, m a 、m b 、m c 、m d All are fitting parameter values ​​of self-discharge rate coefficient; Step 1.3, obtain the semi-empirical formula of battery capacity decay performance, fit the relevant parameters through the battery charge and discharge capacity decay experimental test, and obtain the battery dynamic capacity decay performance model, specifically: In the formula, k Nc is the cycle number coefficient, Nc n is the number of cycles equivalent to 100% charge and discharge cycle, N is the number of experiments, and the average value of the state of charge in a single charge and discharge cycle is defined as SoC a , using the mean square error of the state of charge SoC d Describes the fluctuation amplitude of the state of charge during the cycle, τ is the time constant, where k T are the temperature influence coefficients, T and T ref are the actual value and reference value of the ambient temperature, T a,t and T refa are the actual value and reference value of the internal operating temperature of the VRB, respectively; k1, k2, k3, and k4 are the fitting coefficients of the relationship between capacity attenuation and state of charge.

3. The nested double-layer optimization operation method of the active distribution network all-vanadium liquid flow battery energy storage system taking into account dynamic performance according to claim 1 is characterized in that: Step 2 is to model the dynamic output characteristics of the energy storage battery. The specific model is: Where, SoC t and SoC t-1 They represent the state of charge at time t and time t-1 respectively, E VRB represents the capacity of VRB, γ is the self-discharge rate, α is the capacity decay rate, η c / d represents the charging or discharging efficiency of the all-vanadium liquid flow battery, P VRB Output power of the all-vanadium liquid flow battery energy storage system.

4. The nested double-layer optimization operation method and system of the active distribution network all-vanadium liquid flow battery energy storage system taking into account dynamic performance according to claim 1, characterized in that: The specific process of establishing the comprehensive operation benefit indicator system of active distribution network is as follows: Step 3.1: Consider the abandoned PV as the operating income of the active distribution network and determine the PV consumption: Photovoltaic consumption B ab The calculation is as follows: B ab =k a W VRB In the formula, k a is the price factor, W VRB is the amount of abandoned photovoltaic power generation, where the total amount of abandoned photovoltaic power generation is calculated as follows: Where P wind is the power generation of the wind turbine in the energy storage system, N wind is the installed capacity of wind power, P solar is the photovoltaic power generation in the energy storage system, N solar is the number of photovoltaic installed capacity, P grid_min P is the minimum value of the amount of electricity that the energy storage system purchases from the upper power grid when the active distribution network is running. loss is the network loss, P load is the load power, P VRB is the output power of the energy storage system, Z t It is the characteristic quantity of photovoltaic abandonment state of energy storage system; Step 3.2: Peak shaving and valley filling is the operating benefit of the active distribution network. Calculate the peak shaving and valley filling benefit B up The specific method is as follows: Calculate the peak shaving rate ω, the specific formula is: Where P max is the load peak value during the energy storage operation before peak shaving, P' max It is the load peak value during the energy storage operation after peak shaving; Calculate the number of years N for delaying grid transformation and upgrading based on the peak reduction rate y , the specific calculation formula is as follows: Where, ε is the annual load growth rate; The benefits of delaying the transformation and upgrading obtained by optimizing the energy storage operation are calculated based on the number of years of delaying the transformation and upgrading of the power grid. The formula is as follows: In the formula, C e The investment required for grid transformation and upgrading is r With d r are the interest rate and discount rate respectively; Step 3.3: Calculate the power market arbitrage B arb : In the formula, τ is the total year, Δf is the difference in the operating income of the distribution network in a year, and C p is the real-time electricity price, P sell is the electricity sold by the energy storage system. When the energy storage system sells electricity to the load, P sell >0, when purchasing electricity P sell <0;P grid The energy storage system reduces the amount of electricity purchased from the upper grid by storing photovoltaic power; Step 3.4: Calculate environmental benefits B em : In the above formula, C w,k Charge fees to each unit that emits pollutants; grid,k is the pollutant emission density of the upper power grid; P g It represents the total amount of electricity purchased from the upper power grid that is reduced when the energy storage system is running, M represents the type of pollutant, and P grid It indicates the total amount of electricity purchased from the upper power grid that is reduced every hour when the energy storage system is running.

5. The nested double-layer optimization operation method of the active distribution network all-vanadium liquid flow battery energy storage system taking into account dynamic performance according to claim 1 is characterized in that: The metamodel heuristic method is adopted to construct a nested two-layer optimization operation model with the dynamic performance of all-vanadium liquid flow battery as the inner optimization target and the comprehensive operation benefit of active distribution network as the outer optimization target. The inner and outer objective functions of the nested two-layer optimization operation model are: Inner objective function F1 = max{ω1η c / d -ω2λ-ω3α}; Outer objective function F2 = max{β1B ab +β2B up +β3B arb +β4B em } Among them, ω1, ω2, ω3, β1, β2, β3, and β4 are the weight values ​​of the corresponding decision variables.

6. The nested double-layer optimization operation method of the active distribution network all-vanadium liquid flow battery energy storage system taking into account dynamic performance according to claim 1 is characterized in that: The optimized particle swarm algorithm is used to solve the nested two-layer optimization operation model. The specific process of obtaining the nested two-layer optimization operation scheme of the energy storage system is as follows: Step 5.1: Add constraints to the nested two-layer optimization model, including: For a small distribution network containing photovoltaic power generation and wind power generation, the power balance satisfies the following equation: Among them, P grid is the power transmitted from the upper power grid to this area, P wind The power input to the region for wind power generation in the region, P solar The power input to the area for photovoltaic power generation, P VRB is the power of the energy storage device; 2) Due to the capacity limitation of the transformer, the power P input into this area from the upper level power grid is grid Will be restricted to a certain range, namely: P grid_min ≤P grid ≤P grid_max - Among them, P grid_min Enter the minimum power value of the area for the upper level power grid, P grid_max Input the maximum power of the area to the upper level power grid; 3) SoC for energy storage battery systems t Between maximum state of charge SoC max and minimum state of charge SoC min Between SoC min ≤SoC t ≤SoC max In the energy storage system, SoC is used to reflect the remaining capacity of the energy storage battery. The calculation method is as follows: Among them, E VRB is the rated capacity of the energy storage system battery, P VRB is the real-time power of the energy storage system. VRB When P is greater than 0, the energy storage system discharges. VRB When it is less than 0, the energy storage battery is charged; Step 5.2: Use the improved particle swarm algorithm to solve the nested two-layer optimization operation model. The specific process is as follows: (1) Collect data, including photovoltaic power generation, wind power generation, and power load; (2) Initialize the particle swarm, take the real-time power and charge state of each energy storage as a particle, and assign it an initial position and initial velocity; (3) According to the charge and discharge efficiency, capacity attenuation rate, self-discharge rate and economic benefit index function that need to be solved, the fitness of each particle is calculated by the inner layer F1 and the outer layer objective function F2 of the nested double-layer optimization operation model; according to the fitness value of each particle, the global optimal position is solved; (4) Update the particle position and particle velocity; (5) Determine whether the number of iterations is met. If not, continue to update the particle fitness and position. If so, the calculation ends and the best position, i.e., the optimal solution, is obtained.

Images