Wind and light storage multi-port converter double-layer power smoothing method considering energy storage life loss

Abstract

The invention discloses a wind and light storage multi-port converter double-layer power smoothing method considering energy storage life loss, and belongs to the technical field of wind and light storage power smoothing control. The invention aims to provide a double-layer power smoothing method for a wind-light storage multi-port converter considering energy storage life loss, which comprehensively utilizes load shedding wind power, load shedding photovoltaic and energy storage flexible control to realize integral output power smoothing of a wind-light storage multi-port converter system. According to the method, a linear equivalent energy storage life loss function is constructed, the output power of the wind and light storage multi-port converter system is predicted and controlled by adopting a double-layer model, and smooth overall output power of the wind and light storage multi-port converter system is realized by comprehensively utilizing load shedding wind power, load shedding photovoltaic power and energy storage flexible control; and smooth fluctuation power is optimized and decomposed into energy storage and load shedding wind power and photovoltaic power. Smooth fluctuation power is optimized and decomposed into energy storage and load shedding wind power and photovoltaic power, the charging and discharging depth and frequency of energy storage are reduced, and the service life of the energy storage is prolonged.

Description

Technical Field

[0001] The invention belongs to the technical field of wind-solar-power storage power smoothing control. Background Art

[0002] The wind, solar and storage multi-port converter system is a multi-energy complementary hybrid power generation and conversion device that integrates photovoltaic power generation system, wind power generation and energy storage system. It converts wind energy and light energy into electrical energy and stores the excess electrical energy in the battery energy storage system, thereby reducing wind and light abandonment and reducing the area occupied by wind and light. It uses the complementary characteristics of wind, solar and storage to become a more advanced new energy power generation equipment. With the improvement of environmental awareness and the rapid development of renewable energy, the wind, solar and storage multi-port converter may become an important equipment in the future energy field.

[0003] As the proportion of wind power and photovoltaic installations in the power system increases, their inherent shortcomings become more obvious. First, the output power of wind power and photovoltaics fluctuates greatly. After being connected to the grid, they will bring certain power shocks and fluctuations to the grid, resulting in voltage and frequency fluctuations at the grid access end, threatening the stable operation of the grid; wind turbines and photovoltaic arrays each occupy a large area, but due to the different heights of wind turbines and photovoltaics, they are complementary in installation and can reduce the area occupied. Connecting wind, light, and storage through a multi-port converter can reduce the volatility of output power and alleviate the impact of new energy grid connection on the grid.

[0004] In wind, solar and storage multi-port converters, energy storage charging and discharging can be used to smooth wind and light power fluctuations. However, frequent charging / discharging will reduce the service life of energy storage, resulting in increased operation and maintenance costs. How to maximize the life of energy storage while meeting the output power smoothing requirements is a challenge currently encountered in engineering applications.

[0005] Patent CN117394384A discloses a fuzzy controller based on traditional energy storage, which takes into account the fluctuation trend of the output power of the wind-solar combined power generation system and the parameters of multiple types of energy storage, and controls the output power of the wind-solar combined power generation system based on the actual required power and actual available power of each type of energy storage. However, this invention does not give full play to the operating characteristics of power-type energy storage and energy-type energy storage, ignores the consistency of various types of energy storage, and does not fully consider the cycle life and regulation capacity of various types of energy storage.

[0006] Patent CN116031903A discloses a method of obtaining ultra-short-term power forecast data of wind power and photovoltaics, performing spectrum analysis on the real-time fluctuating power of wind power and photovoltaic systems, calculating the frequency distribution of wind power fluctuations, dynamically correcting the filter cutoff frequency, and improving the smoothness of the overall external power output of the system through coordination between different types of power sources. However, this strategy does not take into account the changes in low-frequency components caused by prediction errors, and does not take into account the coordination with conventional wind and solar power regulation. Therefore, it is impossible to fully realize the overall smooth control of system power and has great application limitations.

[0007] Patent CN118889577A discloses a method for optimizing the active power of a wind-solar-storage station. Based on the actual output active power value of the wind-solar-storage station, the wind-solar-short-term power forecast value and the wind-solar-storage active power control optimization variables, the penalty coefficient of the wind-solar-storage active power control cost is calculated to establish an optimization objective function, and then based on the wind-solar-storage active power control optimization variables, the wind-solar-storage installed capacity is allocated a reasonable proportion to further improve the economic efficiency of system operation. However, the invention does not involve a loss cost optimization strategy for the life of energy storage cycles.

[0008] Although the above method can reduce the volatility of the output power of new energy, the practical use of energy storage is limited by factors such as limited life, energy loss during charging and discharging, and high installation costs. Frequent charging and discharging switching of energy storage will accelerate the loss of cycle life and affect the operating economy of wind, solar and energy storage power generation. However, currently, in the wind and solar power fluctuation smoothing control algorithm, the analysis of energy storage cycle life is rarely designed. In view of the shortcomings of the prior art, the present invention proposes a method for smoothing power fluctuations predicted by a double-layer model of a wind, solar and energy storage multi-port converter to improve the life of energy storage. Summary of the invention

[0009] The purpose of the present invention is to comprehensively utilize load shedding wind power, load shedding photovoltaic power and flexible control of energy storage to achieve overall output power smoothing of a wind-solar-storage multi-port converter system taking into account energy storage life loss to provide a double-layer power smoothing method for a wind-solar-storage multi-port converter.

[0010] The steps of the present invention are: S1. The power model of wind-solar-storage multi-port converter is obtained from the power flow in the wind-solar-storage multi-port converter: P w (k) = P MAP,w (k)+P de,w (k) (1) P pv (k) = P MAP,pv (k)+P de,pv (k) (2) P g (k+1)=P b (k)+P w(k)+P pv (k) (3) C SOE,b (k+1)=C SOE,b (k)-P b (k)×T c / Q b (4) Where P w (k) represents the actual total output power of wind power at time k; P pv (k) represents the actual total photovoltaic output power at time k; P de,w (k) represents the wind power load reduction power at time k; P de,pv (k) represents the photovoltaic load reduction power at time k; P MAP,W (k) represents the theoretical output power of wind power at the time k under the fixed load reduction level mode; P MAP,pv (k) represents the theoretical output power of photovoltaic power in the fixed load reduction level mode at time k; P b (k) represents the charging and discharging power of the energy storage battery at time k; P g (k+1) represents the grid-connected power at time k+1; C SOE,b (k) refers to the energy state of the battery at time k; T c is the control cycle of the energy storage battery; Q b is the storage capacity of the energy storage battery; The load shedding wind and solar adjustable power constraint settings are as follows: -d %P MAP,W ≤P de,W (k)≤d%P MAP,W (5) -d %P MAP,pv ≤P de,pv (k)≤d%P MAP,pv (6) P de (k) = P de,W (k)+P de,pv (k) (7) S2, Linearized energy storage life equivalent cycle loss model: A fourth-order function is used to characterize the relationship between the number of full cycles N and the charge and discharge depth d: N=-3278d 4 -5d 3 +12823d 2 -14122d+5112 (8) The equivalent charge and discharge times of a battery are defined as: Where, d i is the i-th charge and discharge depth, C NTi (di ) indicates that the charge and discharge depth is d i is a full cycle or a half cycle, N(1) is the limit cycle number of the battery, N(d i ) is the charge and discharge depth d i The maximum number of cycles, n i (d i ) indicates that the energy storage battery is at a charge and discharge depth of d i Completing one charge and discharge in 1 hour is equivalent to the number of charge and discharge times under full charge and full discharge conditions; Assume that the battery is charged and discharged m times a day, and the depth of each charge and discharge is d1, d2, ...d m , then the equivalent number of charge and discharge times per day is: Therefore, the operating life of the energy storage battery is: Considering different depths of discharge, the DOD-cycle life loss function of battery energy storage is: The linearized battery energy storage life loss is obtained by piecewise linearization: In the formula, N is the number of segments, and λ is the sequence number of the segment; is the length of the corresponding segment; is the slope of the corresponding segment; S3, double-layer MPC control structure of wind, solar and storage multi-port system: Through the upper model predictive control method, the total smoothed power P is obtained ex (k) control sequence, the total smoothed power P ex The control sequence of (k) is introduced into the lower-level MPC control during the optimization process of the lower-level model predictive control; S4. Two-layer model predictive power smoothing control: minimum exchange power upper layer MPC control method The upper control MPC rolling optimization objective function is: Where P ex is the total smoothed charging and discharging power of energy storage and load shedding wind and solar power; M is the prediction time length; The following constraints must also be met: |P g (k+i+1)-P g (k+i)|≤δ,i=0,1,...,M-1 (15) -P b,max ≤P ex (k+i)≤Pb,max ,i=0,1,...,M-1 (16) P g (k+1)=P ex (k)+P MAP,w (k)+P MAP,pv (k) (17) Where, δ is the allowable value of grid-connected power fluctuation; P b,max is the rated power of the energy storage battery; The power fluctuation of wind and solar power generation is limited by the constraint condition of formula (15), so that the power fluctuation value of wind, solar and energy storage system connected to the grid at time k+i and time k+i+1 is less than ±δMW; S5. Double-layer model predictive power smoothing control: Wind-solar-storage coordinated lower-layer MPC control based on minimum equivalent energy storage cycle loss The objective function of the lower-layer MPC rolling optimization of the double-layer MPC is: Where, L loss(k) is the life loss function of the energy storage battery; L, K are weight coefficients; At the same time, the following constraints are also met: C SOE,min ≤C SOE,b (k+i)≤C SOE,max ,i=1,...,M (19) -P b,max ≤P b (k+i)≤P b,max ,i=0,1,...,M-1 (20) -d%(P MAP,W (k+i)+P MAP,pv (k+i))≤P de (k+i)≤d%(P MAP,W (k+i)+P MAP,pv (k+i)) ,i=0,1,...,M-1 (21)P ex (k+i)=P b (k+i)+P de (k+i) ,i=0,1,...,M-1 (22) At this time, wind power and photovoltaic power are running in the load reduction state, and the lower-level MPC method is used to optimize the solution, and the optimization solution result P of the upper-level MPC is converted into ex (k) is decomposed into energy storage and load shedding wind and solar output power control sequence P b(k) and P de(k) .

[0011] The implementation process of the double-layer MPC power smoothing method of the present invention is as follows: Step 1: Establish a mathematical model of the wind-solar-storage combined system and initialize the parameters at time t=1; Step 2: Update the wind and solar power P at time k MAP,W(k) , P MAP.pv (k), using the wind-solar power and the wind-solar power prediction model to obtain the wind-solar power at time k+1, k+2, …, k+M, and send it to the MPC controller; Step 3: Taking formula (14) as the objective function and considering the constraints of formulas (1)-(7) and (15)-(17), the upper-level MPC method is applied to obtain the total smoothed power P through quadratic programming. ex(k) Control sequence [u(k),u(k+1),…,u(k+M-1)]; Step 4: According to the upper layer solution result P ex(k) The positive or negative power determines the lower layer P ex(k) output range, ensuring that the power directions of the two are consistent. If P ex(k) >0, the output power range of wind and solar load shedding is [0,d%P MAP ]; if P ex(k) <0, the output power range of wind and solar load shedding is [-d%P MAP ,0]; if P ex(k) =0, then the output power of wind and solar power is set to 0; Step 5: Update the energy storage state C at time k SOE,b (k), taking formula (18) as the objective function and considering the constraints of formulas (19)-(22), the lower-level MPC method is applied to the control sequence P solved by the upper level. ex(k) The energy storage power P is obtained by optimizing the quadratic programming. b(k) and load shedding wind and solar power P de(k) The control sequence [u1(k),u2(k),u1(k+1),u2( k +1),…,u1(k+M-1),u2(k+M-1)]. Step 6: The optimization is completed at the current moment, scroll to the next moment t=t+1 and return to step 2 until the time is up.

[0012] The present invention optimizes and decomposes the smoothed fluctuating power into energy storage and load shedding wind power and photovoltaic power, reduces the depth and number of charge and discharge of energy storage, and increases the service life of energy storage. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 It is a multi-port system of wind, solar, storage and power generation and a power flow diagram; Figure 2 This is the double-layer MPC control structure diagram of the present invention Figure 3 Piecewise linear curve of cycle life loss of typical energy storage batteries; Figure 4 The double-layer MPC control flow chart of the present invention; Figure 5 This is the photovoltaic power generation power diagram under cloudy conditions; Figure 6 This is the total power diagram of wind and solar power generation under cloudy conditions; Figure 7 This is the photovoltaic power generation power diagram under sunny sky; Figure 8 This is the total power diagram of wind power generation of photovoltaic power in sunny days; Fig. 9 It is a global diagram comparing the grid-connected power under different control methods at a 5% load reduction level on a sunny day; Fig.10 This is a comparison chart of grid-connected power fluctuations under different control methods at a 5% load reduction level on a sunny day; Fig.11 This is the load shedding wind and solar output diagram under 5% load shedding level in sunny weather; Fig.12 It is a global diagram comparing the energy storage output power under different control methods at a 5% load reduction level on a sunny day; Fig.13 This is a comparison chart of energy storage SOC under different control methods at a 5% load reduction level on a sunny day; Fig.14 This is a comparison chart of grid-connected power fluctuations at a 5% load reduction level on cloudy days; Fig.15 It is the energy storage SOC under different control methods at 5% load reduction level on cloudy days. DETAILED DESCRIPTION

[0014] The present invention constructs a linear equivalent energy storage life loss function, adopts a double-layer model to predict and control the output power of a wind-solar-storage multi-port converter system, and comprehensively utilizes load shedding wind power, load shedding photovoltaics and energy storage flexible control to achieve smooth overall output power of the wind-solar-storage multi-port converter system, and optimizes the decomposition of the smoothed fluctuating power into energy storage and load shedding wind power and photovoltaics, thereby reducing the depth and number of charge and discharge of energy storage and improving the service life of energy storage.

[0015] The objective of the present invention is achieved by the following technical solutions: 1) Wind and solar power conversion DC port load reduction power control Wind and solar power units generally use the maximum power point tracking (MPPT) method to access the power grid, and do not have flexible power control capabilities. In this invention, the wind and solar power generation port is changed from maximum power point tracking control to load reduction control to obtain some backup power. Load reduction control can be achieved by changing the operating voltage of the wind turbine and photovoltaic array. Load reduction control enables wind power and photovoltaic to have a certain degree of flexible power control capabilities.

[0016] 2) Energy storage DC port fluctuation smoothing control The converter of the energy storage system of the present invention adopts a constant power control strategy with a power smoothing function. The bottom layer control adopts a dual-loop control strategy of power outer loop and current inner loop, and sets the reference value P ref With the measured value P e,b After the difference is made, it passes through the PI controller and then the limiting link to obtain the current reference value, which is then compared with the actual current value I b After making the difference, the input signal of the converter is obtained through the PI controller.

[0017] Reference value P ref In order to smooth the power command, a double-layer power smoothing method is proposed. On the basis of power control, wind and light power smoothing functions are provided at the same time.

[0018] 3) Energy storage life equivalent cycle loss function Energy storage has the characteristics of bidirectional power transmission of charge / discharge, fast control speed, and good power controllability, which can be used to smooth the fluctuation of wind and solar power output power. However, frequent charging and discharging will greatly affect the life of energy storage and increase operating costs. Related studies have shown that battery life is related to factors such as temperature, charging method, charging state, peak current, and cycle state.

[0019] To simplify the calculation, only the effects of charge and discharge times and depth of discharge (DOD) on the battery cycle life are considered. The larger the DOD, the shorter the service life, and the smaller the DOD, the longer the service life. The DOD-cycle life curve of battery energy storage is relatively complex and difficult to embed into the optimization model analysis using explicit expressions. The present invention uses the number of full charge and discharge trial cycles of energy storage to define the equivalent number of cycles at different charge and discharge depths, and uses piecewise linearization to obtain the linearized battery energy storage life equivalent cycle loss function.

[0020] The energy storage life loss model can be embedded into the lower control target model in the double-layer predictive control optimization control, making it possible to improve the energy storage life control.

[0021] 4) Double-layer model predictive control (MPC) power smoothing control: Upper-layer MPC control The power smoothing algorithm of the present invention adopts double-layer MPC control. The upper-layer MPC minimizes the total smooth exchange power. That is, the requirement of wind power and photovoltaic output power smoothing is met at the cost of the minimum exchange power.

[0022] The upper-level MPC control input includes the wind and solar power P at time k MAP,W(k) , P MAP.pv(k) , using the previous wind and solar power and wind and solar power prediction model to obtain the wind and solar power at time k+1, k+2, ..., k+M, the output is the total wind and solar power P of energy storage and load reduction obtained by quadratic programming optimization ex (k) is the control sequence [u(k),u(k+1),…,u(k+M-1)].

[0023] The upper-level control optimizes the system with the minimum total smoothed power as the objective function on the basis of satisfying the grid-connected fluctuation constraint and power constraint. The upper-level MPC control sets constraints to ensure that the wind power and photovoltaic power with large fluctuations originally meet the grid-connected requirements and ensure the minimum total output of energy storage and load shedding.

[0024] 5) Two-layer model predictive control (MPC) power smoothing control: The lower layer MPC control input includes the energy state C of the energy storage at time k. SOE,b (k), and the upper-level MPC optimization result P ex (k) control sequence. The output is the energy storage power P b (k) and load shedding wind power and photovoltaic power P de (k) is the control sequence of [u1(k), u2(k), u1(k+1), u2(k+1), …, u1(k+M-1), u2(k+M-1)].

[0025] The lower-level control, based on satisfying the energy storage SOE state constraints, energy storage output power constraints, and wind power and photovoltaic load reduction power constraints, utilizes the load reduction wind and photovoltaic power and energy storage to coordinate output, with the minimization of the energy state change trend of the energy storage battery and the minimization of the equivalent energy storage life loss as sub-objective functions. It can effectively reduce the number of charge and discharge cycles and discharge depth of the energy storage battery and improve the service life of the energy storage.

[0026] The present invention will be further described below in conjunction with the accompanying drawings and implementation examples: 1. Wind, solar and energy storage multi-port converter power model Depend on Figure 1 The power flow in the wind-solar-storage multi-port converter can be obtained: P w (k) = P MAP,w(k)+P de,w (k) (1) P pv (k) = P MAP,pv (k)+P de,pv (k) (2) P g (k+1)=P b (k)+P w (k)+P pv (k) (3) C SOE,b (k+1)=C SOE,b (k)-P b (k)×T c / Q b (4) Where P w (k) represents the actual total output power of wind power at time k; P pv (k) represents the actual total photovoltaic output power at time k; P de,w (k) represents the wind power load reduction power at time k; P de,pv (k) represents the photovoltaic load reduction power at time k; P MAP,W (k) represents the theoretical output power of wind power at the time k under the fixed load reduction level mode; P MAP,pv (k) represents the theoretical output power of photovoltaic power in the fixed load reduction level mode at time k; P b (k) represents the charging and discharging power of the energy storage battery at time k; P g (k+1) represents the grid-connected power at time k+1; C SOE,b (k) refers to the energy state of the battery at time k; T c is the control cycle of the energy storage battery; Q b is the storage capacity of the energy storage battery.

[0027] The basic idea of ​​using energy storage to smooth out wind and solar power fluctuations is to absorb and compensate for the fluctuation component in wind and solar power through the charging and discharging of energy storage. The load shedding wind and solar adjustable power constraint settings are as follows: -d %P MAP,W ≤P de,W (k)≤d%P MAP,W (5) -d %P MAP,pv ≤P de,pv (k)≤d%P MAP,pv (6) P de (k) = P de,W (k)+P de,pv (k) (7)

[0028] from Figure 1It can be seen that the original data values ​​of wind and photovoltaic power generation and the state feedback values ​​of energy storage are input into the energy management system. After being calculated and processed by the optimization algorithm in the energy management system, the power control signals are output to the load reduction wind power port, load reduction photovoltaic port and energy storage port respectively.

[0029] 2. Linearized energy storage life equivalent cycle loss model The depth of discharge (DOD) of energy storage batteries is an important factor affecting the operating life of energy storage batteries. The larger the DOD, the shorter the service life, and the smaller the DOD, the longer the service life. A fourth-order function can be used to characterize the relationship between the number of full cycles N and the depth of discharge d: N=-3278d 4 -5d 3 +12823d 2 -14122d+5112 (8)

[0030] Due to the randomness of power fluctuations, the depth of each charge and discharge is different, making it difficult to count the number of cycles. The number of cycles under different DOD is equivalent to the number of cycles under full charge and discharge. The equivalent charge and discharge times of the battery are defined as: Where, d i is the i-th charge and discharge depth, C NTi (d i ) indicates that the charge and discharge depth is d i is a full cycle or a half cycle, the full cycle is 1, the half cycle is 0.5, N(1) is the maximum cycle number of the battery, that is, the maximum number of cycles of full charge and discharge, N(d i ) is the charge and discharge depth d i The maximum number of cycles, n i (d i ) indicates that the energy storage battery is at a charge and discharge depth of d i Completing one charge and discharge in 1 hour is equivalent to the number of charge and discharge times under full charge and full discharge conditions.

[0031] Assume that the battery is charged and discharged m times a day, and the depth of each charge and discharge is d1, d2, ...d m , then the equivalent number of charge and discharge times per day is:

[0032] Therefore, the calculation formula for the operating life of the energy storage battery is:

[0033] Considering different depths of discharge, the DOD-cycle life loss function of battery energy storage is:

[0034] However, since the DOD-cycle life curve of battery energy storage is relatively complex, it is difficult to embed the optimization model analysis with explicit expressions. The linearized battery energy storage life loss is obtained by piecewise linearization, as shown in formula (13): In formula (13), N is the number of segments, and the present invention takes the number of segments as 5; λ is the sequence number of the segment; is the length of the corresponding segment; To correspond to the segmented slopes, the battery energy storage life loss model embedded in the optimization model makes the analysis more convenient.

[0035] The energy storage life loss curves of two types of energy storage batteries NeoVoltaNV14 (LiFePO4 battery and Moura12MF220 (lead-acid battery) obtained by piecewise linear fitting using the method of the present invention are shown in Figure 2 shown.

[0036] 3. Double-layer MPC control structure of wind, solar and storage multi-port system The block diagram of the double-layer MPC power smoothing control method mentioned in the present invention is as follows: Figure 3 shown. Figure 3 The wind power and photovoltaic power generation power forecast series are based on wind power and photovoltaic power forecasting technology. Figure 3 In the above example, the total smoothed power P is obtained by using the upper model predictive control method. ex (k) The total smoothed power P ex The control sequence of (k) is introduced into the lower-level MPC control in the optimization process of the lower-level model predictive control, and the load-shedding wind power, photovoltaic power and energy storage are used to coordinate the output to minimize the energy state change trend of the energy storage battery and the energy storage life loss.

[0037] 4. Two-layer model predictive power smoothing control: minimum exchange power upper layer MPC control method The upper control MPC rolling optimization objective function is: Where P ex is the total smoothed charging and discharging power of energy storage and load shedding wind and solar power; M is the prediction time length, which is also the length of the control instruction sequence.

[0038] When using formula (14) as the optimization objective function, the following constraints must also be met: |P g (k+i+1)-P g (k+i)|≤δ,i=0,1,...,M-1 (15) -Pb,max ≤P ex (k+i)≤P b,max ,i=0,1,...,M-1 (16) P g (k+1)=P ex (k)+P MAP,w (k)+P MAP,pv (k) (17) Where, δ is the allowable value of grid-connected power fluctuation; P b,max is the rated power of the energy storage battery.

[0039] The constraint condition of formula (15) is used to limit the fluctuation of wind and solar power generation, so that the grid-connected power fluctuation value of the wind, solar and energy storage system at time k+i and time k+i+1 is less than ±δMW, thereby limiting the grid-connected wind and solar power fluctuation within the threshold.

[0040] The upper control of the two-layer MPC optimizes and solves the control sequence P of the total power of energy storage and load shedding to minimize the total power of wind and solar power on the basis of ensuring that the grid-connected fluctuation is below the constraint threshold. ex (k) At this stage, the operation of energy storage and load reduction wind and solar power is temporarily not taken into account.

[0041] 5. Double-layer model predictive power smoothing control: Wind-solar-storage coordinated lower-layer MPC control based on minimum equivalent energy storage cycle loss: The lower-level control, based on satisfying the SOE state constraints of energy storage, energy storage output power constraints, and wind and solar load reduction power constraints, uses load reduction wind and solar power and energy storage to coordinate output, with the minimum energy state change trend of energy storage batteries and the minimum loss of equivalent energy storage cycle life as the objective function. By setting the objective function, the number of charge and discharge cycles and the depth of discharge of energy storage batteries can be effectively reduced, and the energy storage life can be improved.

[0042] The objective function of the lower control MPC rolling optimization of the double-layer MPC is: Where, L loss(k) is the life loss function of the energy storage battery; L, K are weight coefficients.

[0043] At the same time, the following constraints must be met: C SOE,min ≤C SOE,b (k+i)≤C SOE,max ,i=1,...,M (19) -P b,max ≤P b (k+i)≤P b,max ,i=0,1,...,M-1 (20) -d%(PMAP,W (k+i)+P MAP,pv (k+i))≤P de (k+i)≤d%(P MAP,W (k+i)+P MAP,pv (k+i))(21),i=0,1,...,M-1 P ex (k+i)=P b (k+i)+P de (k+i),i=0,1,...,M-1 (22)

[0044] Pay attention to details: P should be kept here de (k+i) and P ex The power direction of (k+i) is consistent, thereby reducing the output of energy storage and not causing waste of energy storage output.

[0045] At this time, wind power and photovoltaic power are running in the load reduction state, and the lower-level MPC method is used to optimize the solution, and the optimization solution result P of the upper-level MPC is converted into ex (k) is decomposed into energy storage and load shedding wind and solar output power control sequence P b(k) and P de(k) .

[0046] The present invention introduces the energy storage power change rate and energy storage life loss function into the lower-layer model predictive control optimization process through the objective function (18), on the premise that the total output of energy storage and load-shedding wind and solar power is minimized through the upper-layer model predictive control rolling optimization. The load-shedding wind and solar power and energy storage are used to coordinate the output, so that the energy storage power change rate tends to the minimum and the energy storage life loss is minimized, effectively reducing the number of frequent charge and discharge cycles and the discharge depth of the energy storage, and realizing the protection of the energy storage and the extension of its life.

[0047] 6. Execution process of the double-layer MPC power smoothing method of the present invention: The flow chart of the double-layer MPC control method mentioned in the present invention is as follows: Figure 4 As shown, the specific process is as follows: Step 1: Establish a mathematical model of the wind-solar-storage combined system and initialize the parameters at time t=1.

[0048] Step 2: Update the wind and solar power P at time k MAP,W(k) , P MAP.pv (k) The wind and solar power at time k+1, k+2, ..., k+M is obtained by using the previous wind and solar power and the wind and solar power prediction model, and is sent to the MPC controller.

[0049] Step 3: Taking formula (14) as the objective function and considering the constraints of formulas (1)-(7) and (15)-(17), the upper-level MPC method is applied to obtain the total smoothed power P through quadratic programming. ex(k) Control sequence [u(k),u(k+1),…,u(k+M-1)].

[0050] Step 4: According to the upper layer solution result P ex(k) The positive or negative power determines the lower layer P ex(k) output range, ensuring that the power directions of the two are consistent. ex(k) >0, the output power range of wind and solar load shedding is [0,d%P MAP ]; if P ex(k) <0, the output power range of wind and solar load shedding is [-d%P MAP ,0]; if P ex(k) =0, then the output power of the load-shedding wind and solar power is set to 0.

[0051] Step 5: Update the energy storage state C at time k SOE,b (k), taking formula (18) as the objective function and considering the constraints of formulas (19)-(22), the lower-level MPC method is applied to the control sequence P solved by the upper level. ex(k) The energy storage power P is obtained by optimizing the quadratic programming. b(k) and load shedding wind and solar power P de(k) The control sequence [u1(k),u2(k),u1(k+1),u2( k +1),…,u1(k+M-1),u2(k+M-1)].

[0052] Step 6: The optimization is completed at the current moment, scroll to the next moment t=t+1 and return to step 2 until the time is up.

[0053] Experimental simulation to evaluate the invented method The present invention selects the measured data of wind and photovoltaic power on a typical day in a certain area for analysis and research, with a sampling interval of 1 minute. Table 1 shows the parameter settings related to the simulation experiment model.

[0054] Table 1 Simulation parameters

[0055] In order to demonstrate the superiority of the invented method, the present invention sets up three groups of different methods for comparison.

[0056] Method 1: Single-layer MPC method with the goal of minimizing energy storage output. The single-layer optimization control method with the goal of minimizing energy storage output ignores the damage to the energy storage life caused by the frequent charge and discharge cycle switching and excessive discharge depth.

[0057] Method 2: Wind and solar load reduction operation, the upper layer is aimed at the minimum output, and the lower layer is a double-layer MPC method with the minimum change in energy storage capacity. The minimum change in energy storage capacity can reduce the number of charge and discharge switching to a certain extent and improve the life of energy storage. However, it does not consider the impact of charge and discharge switching times and the deep integration of charge and discharge cycles on the life loss of energy storage.

[0058] Method 3: The present invention considers a double-layer MPC method for improving energy storage life.

[0059] It should be noted that method 1 is the upper-level MPC method of method 3, so some simulation curves will be repeated with the results of method 3 and will not be repeated. Methods 2 and 3 are both double-layer MPC methods. The goal of method 2 is to minimize the change in energy storage capacity, and the goal of method 3 is to minimize the equivalent cycle loss life of energy storage.

[0060] Two typical weather conditions, a sunny day and a cloudy day, were selected, and the load reduction level for wind power and photovoltaic power was selected as 5%. Figure 5 This is the photovoltaic power generation diagram under sunny sky. Figure 6 This is a superimposed graph of wind and photovoltaic power generation on a sunny day. Figure 7 is the photovoltaic power generation curve under cloudy conditions, Figure 8 It is the superposition curve of wind and photovoltaic power generation on a cloudy day.

[0061] The algorithm simulation experiment results are as follows Figures 9 to 15 shown.

[0062] Fig. 9 This is the power curve of the grid connection point under sunny weather. Fig.10 is the grid-connected power variation curve. Fig. 9 , 10 It can be seen that: (1) When wind and solar power are directly connected to the grid, the grid power fluctuation value is large, and most of the time it does not meet the grid power fluctuation requirements, which will cause a large impact on the power grid; (2) Method 1, Method 2 and Method 3 can all limit the grid power fluctuation within the ideal range and meet the allowable grid power fluctuation requirements.

[0063] Fig.11 This is the load reduction wind and solar output curve under 5% load reduction level in sunny weather. Fig.12 This is the comparison curve of energy storage output power under different control methods at 5% load reduction level on sunny days. Fig.13 This is the energy storage SOC comparison curve under different control methods at a 5% load reduction level on a sunny day.

[0064] Depend on Figures 11 to 13It can be seen that: (1) Method 1 causes the energy storage battery to frequently change its charge and discharge state; (2) Method 2 and Method 3 of the present invention reduce the number of energy storage charge and discharge switching and the depth of discharge by reducing the output of wind and solar power with limited capacity. (3) Especially in Method 3 of the present invention, the charge and discharge curve of the energy storage battery is changed from the original frequent charge and discharge to a relatively gentle slow charge and discharge curve, and the time interval for switching the energy storage charge and discharge state is significantly extended. In some time periods, a step-like change trend will appear, which reduces the number of charge and discharge switching and improves the energy storage life.

[0065] The equivalent energy storage cycle number is calculated by the rain-mile counting method, and the daily equivalent full cycle number and equivalent energy storage cycle life are calculated by formulas (10) and (11), respectively, as shown in Table 2.

[0066] Table 2 Comparison of energy storage life evaluation indicators under sunny weather conditions

[0067] It can be seen from Table 2 that compared with Method 1 and Method 2, the double-layer MPC method of the present invention taking into account the energy storage life loss function changes the number of charge and discharge changes of the energy storage battery N cd Significantly reduced to 82 times; equivalent energy storage life T d Got promoted to 1607 days.

[0068] Fig.14 This is the comparison curve of grid-connected power fluctuation under different control methods on cloudy days. Fig.15 Energy storage SOC under different control methods in cloudy weather. Table 3 below shows the daily equivalent cycle number and equivalent cycle life under cloudy weather conditions.

[0069] Table 3 Comparison of energy storage life evaluation indicators under cloudy conditions

[0070] It can be seen from Table 3 that under the 5% load reduction level on cloudy days, on the basis of reducing the grid-connected fluctuating power, compared with methods 1 and 2, the method 3 of the present invention changes the number of times N of energy storage battery charging and discharging. cd From down to 113 times, the equivalent energy storage cycle life T d It takes 1565 days from promotion to upgrade.

Claims

1. A double-layer power smoothing method for wind, solar and storage multi-port converters taking into account energy storage life loss, characterized in that: The steps are: S1. Wind, solar and energy storage multi-port converter power model According to the power flow in the wind-solar-storage multi-port converter: P w (k)=P MAP,w (k)+P de,w (k) (1) P pv (k)=P MAP,pv (k)+P de,pv (k) (2) P g (k+1)=P b (k)+P w (k)+P pv (k) (3) C SOE,b (k+1)=C SOE,b (k)-P b (k)×T c / Q b (4) Where P w (k) represents the actual total output power of wind power at time k; P pv (k) represents the actual total photovoltaic output power at time k; P de,w (k) represents the wind power load reduction power at time k; P de,pv (k) represents the photovoltaic load reduction power at time k; P MAP,W (k) represents the theoretical output power of wind power at the time k under the fixed load reduction level mode; P MAP,pv (k) represents the theoretical output power of photovoltaic power in the fixed load reduction level mode at time k; P b (k) represents the charging and discharging power of the energy storage battery at time k; P g (k+1) represents the grid-connected power at time k+1; C SOE,b (k) refers to the energy state of the battery at time k; T c is the control cycle of the energy storage battery; Q b is the storage capacity of the energy storage battery; The load shedding wind and solar adjustable power constraint settings are as follows: -d%P MAP,W ≤P de,W (k)≤d%P MAP,W (5) -d%P MAP,pv ≤P de,pv (k)≤d%P MAP,pv (6) P de (k)=P de,W (k)+P de,pv (k) (7) S2, Linearized energy storage life equivalent cycle loss model: A fourth-order function is used to characterize the relationship between the number of full cycles N and the charge and discharge depth d: N=-3278d 4 -5d 3 +12823d 2 -14122d+5112 (8) The equivalent charge and discharge times of a battery are defined as: Where, d i is the i-th charge and discharge depth, C NTi (d i ) indicates that the charge and discharge depth is d i is a full cycle or a half cycle, N(1) is the limit cycle number of the battery, N(d i ) is the charge and discharge depth d i The maximum number of cycles, n i (d i ) indicates that the energy storage battery is at a charge and discharge depth of d i Completing one charge and discharge in 1 hour is equivalent to the number of charge and discharge times under full charge and full discharge conditions; Assume that the battery is charged and discharged m times a day, and the depth of each charge and discharge is d1, d2, ...d m , then the equivalent number of charge and discharge times per day is: Therefore, the operating life of the energy storage battery is: Considering different depths of discharge, the DOD-cycle life loss function of battery energy storage is: The linearized battery energy storage life loss is obtained by piecewise linearization: In the formula, N is the number of segments, and λ is the sequence number of the segment; is the length of the corresponding segment; is the slope of the corresponding segment; S3, double-layer MPC control structure of wind, solar and storage multi-port system: Through the upper model predictive control method, the total smoothed power P is obtained ex (k) control sequence, the total smoothed power P ex The control sequence of (k) is introduced into the lower-level MPC control during the optimization process of the lower-level model predictive control; S4. Two-layer model predictive power smoothing control: minimum exchange power upper layer MPC control method The upper control MPC rolling optimization objective function is: Where P ex is the total smoothed charging and discharging power of energy storage and load shedding wind and solar power; M is the prediction time length; The following constraints must also be met: |P g (k+i+1)-P g (k+i)|≤δ,i=0,1,...,M-1 (15) -P b,max ≤P ex (k+i)≤P b,max ,i=0,1,...,M-1 (16) P g (k+1)=P ex (k)+P MAP,w (k)+P MAP,pv (k) (17) Where, δ is the allowable value of grid-connected power fluctuation; P b,max is the rated power of the energy storage battery; The power fluctuation of wind and solar power generation is limited by the constraint condition of formula (15), so that the power fluctuation value of wind, solar and energy storage system connected to the grid at time k+i and time k+i+1 is less than ±δMW; S5. Double-layer model predictive power smoothing control: Wind-solar-storage coordinated lower-layer MPC control based on minimum equivalent energy storage cycle loss The objective function of the lower-layer MPC rolling optimization of the double-layer MPC is: Where, L loss(k) is the life loss function of the energy storage battery; L, K are weight coefficients; At the same time, the following constraints are also met: C SOE,min ≤C SOE,b (k+i)≤C SOE,max ,i=1,...,M (19) -P b,max ≤P b (k+i)≤P b,max ,i=0,1,…,M-1 (20) -d%(P MAP,W (k+i)+P MAP,pv (k+i))≤P de (k+i)≤d%(P MAP,W (k+i)+P MAP,pv (k+i)), i=0,1,…,M-1 (21)P ex (k+i)=P b (k+i)+P de (k+i), i=0,1,…,M-1 (22) At this time, wind power and photovoltaic power are running in the load reduction state, and the lower-level MPC method is used to optimize the solution, and the optimization solution result P of the upper-level MPC is converted into ex (k) is decomposed into energy storage and load shedding wind and solar output power control sequence P b(k) and P de(k) .

2. The double-layer power smoothing method for wind-solar-storage multi-port converter taking into account energy storage life loss according to claim 1 is characterized in that: The execution process of the two-layer MPC power smoothing method is as follows: Step 1: Establish a mathematical model of the wind-solar-storage combined system and initialize the parameters at time t=1; Step 2: Update the wind and solar power P at time k MAP,W(k) , P MAP.pv (k), using the wind-solar power and the wind-solar power prediction model to obtain the wind-solar power at time k+1, k+2, …, k+M, and send it to the MPC controller; Step 3: Taking formula (14) as the objective function and considering the constraints of formulas (1)-(7) and (15)-(17), the upper-level MPC method is applied to obtain the total smoothed power P through quadratic programming. ex(k) Control sequence [u(k),u(k+1),…,u(k+M-1)]; Step 4: According to the upper layer solution result P ex(k) The positive or negative power determines the lower layer P ex(k) output range, ensuring that the power directions of the two are consistent. If P ex(k) >0, the output power range of wind and solar load shedding is [0,d%P MAP ]; if P ex(k) <0, the output power range of wind and solar load shedding is [-d%P MAP ,0]; if P ex(k) =0, then the output power of wind and solar power is set to 0; Step 5: Update the energy storage state C at time k SOE,b (k), taking formula (18) as the objective function and considering the constraints of formulas (19)-(22), the lower-level MPC method is applied to the control sequence P solved by the upper level. ex(k) The energy storage power P is obtained by optimizing the quadratic programming. b(k) and load shedding wind and solar power P de(k) The control sequence [u1(k),u2(k),u1(k+1),u2( k +1),…,u1(k+M-1),u2(k+M-1)]. Step 6: The optimization is completed at the current moment, scroll to the next moment t=t+1 and return to step 2 until the time is up.

Images