A double-layer hybrid energy storage capacity configuration method for a wind storage combined power generation system

Abstract

The application belongs to the technical field of wind storage combined power generation system, and particularly relates to a double-layer hybrid energy storage capacity configuration method for a wind storage combined power generation system. The method comprises the following steps: step 1, decomposing the hybrid energy storage target power by using MPC; step 2, establishing a hybrid energy storage system capacity configuration model based on the whole life cycle cost; and step 3, solving the hybrid energy storage capacity configuration model by using a particle swarm algorithm, and dynamically adjusting the MPC weight parameters based on the configuration results. The inner layer adopts MPC to dynamically and optimally decompose the given hybrid energy storage total demand power into real-time output instructions of the battery and the super capacitor. The outer layer quantifies the battery life loss by using the rain flow counting method and the equivalent cycle life method based on the battery power sequence output by the MPC, establishes a whole life cycle refined cost model, and globally optimizes the rated capacity and power of the energy storage system by using the particle swarm algorithm.

Description

Technical Field

[0001] This invention relates to the field of wind-storage combined power generation system technology, specifically to a method for configuring a two-layer hybrid energy storage capacity for a wind-storage combined power generation system. Background Technology

[0002] Wind power is a clean and green energy source, and the continuous development of wind resources is an important means for the global energy transition to sustainable development. However, the randomness and volatility of large-scale wind power grid connection pose multiple challenges to the stable operation of the power system. In recent years, hybrid energy storage systems have been widely used to mitigate wind power fluctuations due to their flexible operating characteristics, and extensive research has been conducted on their control methods. However, in practical engineering applications, the operation control of energy storage systems is closely related to capacity configuration. Therefore, integrating dynamic power allocation and capacity optimization configuration into an iterative feedback mechanism is of great significance in this scenario. Summary of the Invention

[0003] The purpose of this section is to outline some aspects of the embodiments of the present invention and to briefly describe some preferred embodiments. Simplifications or omissions may be made in this section, as well as in the abstract and title of this application, to avoid obscuring the purpose of these documents; however, such simplifications or omissions should not be construed as limiting the scope of the invention.

[0004] To address the aforementioned technical problems, according to one aspect of the present invention, the present invention provides the following technical solution: a method for configuring a two-layer hybrid energy storage capacity for a wind-storage combined power generation system, comprising the following steps:

[0005] Step 1: Decompose the target power of hybrid energy storage using MPC: Establish an MPC weighted quadratic objective function, construct a power command allocation model based on MPC, and realize the effective decomposition of the target power of hybrid energy storage;

[0006] Step 2: Calculate the depth of discharge using the rainflow counting method and the battery life using the equivalent cycle life method, and establish a capacity configuration model for the hybrid energy storage system based on the total life cycle cost.

[0007] Step 3: Use the particle swarm optimization algorithm to solve the hybrid energy storage capacity configuration model, and dynamically adjust the MPC weight parameters based on the configuration results.

[0008] As a preferred embodiment of the dual-layer hybrid energy storage capacity configuration method for wind-storage combined power generation system described in this invention, the specific steps for establishing the MPC weighted quadratic objective function in step 1 are as follows: the objective function is set as a multi-objective weighted quadratic function:

[0009]

[0010] Where: N p For prediction in the time domain; Let k be the power that the lithium-ion battery needs to consume at time k; Let k be the power that the supercapacitor needs to absorb at time k. Let Q1 be the target power of hybrid energy storage at time k; Q2 be the weight of the power tracking accuracy of target 1; Q3 be the weight of the battery usage cost of target 2; Q4 be the weight of the supercapacitor usage cost of target 3; and Q5 be the weight of the power smoothing term of target 4. Let k be the change in the power required to be consumed by the lithium-ion battery at time k compared to time k-1. This represents the change in the power that the supercapacitor needs to absorb at time k compared to time k-1.

[0011] As a preferred embodiment of the dual-layer hybrid energy storage capacity configuration method for a wind-storage combined power generation system described in this invention, the specific steps in step 1 of constructing the MPC-based power command allocation model are as follows: The discrete-space state equation of the system is as follows:

[0012]

[0013] In the formula: Let A be the predicted value of the state variable at time k+1; A is the state matrix, describing how the state variable transitions from time k to time k+1. Let B1 be the state variable at time k; B1 is the input matrix, describing how the control variables affect the changes in the state variables. B1 represents the control variable at time k; B2 is the disturbance matrix, describing the influence of the disturbance variable on the state variable. Let k be the perturbation variable at time k;

[0014] Further derivation yields the predicted state variables at time k+2:

[0015]

[0016] The space state equation simplifies to:

[0017]

[0018] Based on the equality constraints in the power allocation process of the hybrid energy storage system, a power allocation space state model based on the MPC algorithm is obtained:

[0019]

[0020] In the formula: This represents the state of charge of the lithium-ion battery at time k+1. This represents the state of charge of the lithium-ion battery at time k. This represents the state of charge of the supercapacitor at time k+1. This represents the state of charge of the supercapacitor at time k; , The average efficiency of batteries and supercapacitors; , Rated capacity for batteries and supercapacitors;

[0021] Will and As a system state variable:

[0022]

[0023] Will and As a control variable of the system:

[0024]

[0025] Combining the two, we obtain the complete discrete state-space equation:

[0026] .

[0027] As a preferred embodiment of the dual-layer hybrid energy storage capacity configuration method for wind-storage combined power generation system described in this invention, the constraints considered in the power command allocation model are mainly as follows:

[0028] Equality constraints:

[0029]

[0030] SOC Constraints: To prevent overcharging and over-discharging of energy storage devices during wind power fluctuation smoothing, the state of charge (SOC) of energy storage should comply with the following constraints:

[0031]

[0032]

[0033] In the formula: The battery is in its state of charge. This is the state of charge of a supercapacitor;

[0034] Power constraints:

[0035]

[0036]

[0037] In the formula: This refers to the lower limit of battery charging and discharging power. This refers to the upper limit of battery charging and discharging power. This refers to the lower limit of the charging and discharging power of supercapacitors. This represents the upper limit of the charging and discharging power of the supercapacitor.

[0038] As a preferred embodiment of the dual-layer hybrid energy storage capacity configuration method for a wind-storage combined power generation system described in this invention, the specific steps of calculating the depth of discharge using the rainflow counting method in step 2 are as follows: a simplified rainflow counting method is used to obtain the DOD of the battery in each operation, and the cycle parameters obtained by the rainflow counting method are substituted into the battery capacity decay model to obtain the battery life.

[0039] As a preferred embodiment of the dual-layer hybrid energy storage capacity configuration method for wind-storage combined power generation system described in this invention, the specific steps for calculating battery life using the equivalent cycle life method in step 2 are as follows: considering only the impact of DOD level on battery capacity decay, an exponential decay function is used for fitting, and the resulting double exponential decay function characterizes the relationship between depth of discharge and cycle life.

[0040]

[0041] In the formula: This refers to the number of battery cycles. The depth of charge and discharge of the battery;

[0042] The equivalent cycle life method is used to convert the charge-discharge cycles with different depths of discharge experienced by the battery in actual operation into standard cycle numbers at 100% depth of discharge, according to the principle of equivalent life loss. The expression for the equivalent cycle number is then obtained through calculation:

[0043]

[0044] In the formula: This represents the equivalent number of iterations. The number of cycles; This represents the number of cycles the battery corresponds to when the depth of discharge is 1. For the battery at a depth of discharge of The number of loops corresponding to the time;

[0045]

[0046] In the formula: T represents the battery life loss.

[0047] As a preferred embodiment of the dual-layer hybrid energy storage capacity configuration method for wind-storage combined power generation systems described in this invention, the specific steps in step 2 of establishing a hybrid energy storage system capacity configuration model based on total life-cycle cost are as follows: taking the minimum total life-cycle cost as the objective function, comprehensively covering initial investment purchase, supporting auxiliary equipment, long-term operation and maintenance, equipment replacement, and final scrapping and recycling revenue, i.e.:

[0048]

[0049] In the formula: For investment and installation costs; To cover auxiliary costs; For operation and maintenance costs; For replacement costs; For the revenue from scrap recycling;

[0050] Investment and installation costs:

[0051]

[0052]

[0053]

[0054]

[0055]

[0056] In the formula: , These are the investment costs for batteries and supercapacitors, respectively. , The installation costs are for batteries and supercapacitors, respectively. , These are the unit power configuration costs for batteries and supercapacitors, respectively. , These are the unit capacity configuration costs for batteries and supercapacitors, respectively. , The rated power is configured for the battery and the supercapacitor, respectively; , These are the rated capacities for the battery and supercapacitor configurations, respectively. , These are the installation coefficients for batteries and supercapacitors, respectively.

[0057] Ancillary costs:

[0058]

[0059]

[0060]

[0061] In the formula: , These are the auxiliary costs for batteries and supercapacitors, respectively. , These are the cost coefficients for auxiliary equipment for batteries and supercapacitors, respectively.

[0062] Operation and maintenance costs:

[0063]

[0064]

[0065]

[0066]

[0067] In the formula: , These are the operating and maintenance costs of batteries and supercapacitors, respectively. , These are the operation and maintenance cost coefficients for batteries and supercapacitors, respectively; PWF is the present value factor of an equal annuity; T is the project analysis period. The benchmark discount rate;

[0068] Replacement cost:

[0069]

[0070]

[0071] In the formula: n is the number of battery replacements during the project cycle; denoted as , where is the battery's lifespan; represents the i-th replacement, with replacement cost taking into account cost reductions due to technological advancements.

[0072] Revenue from scrap recycling:

[0073]

[0074]

[0075]

[0076]

[0077]

[0078] In the formula: , The residual value gains for batteries and supercapacitors, respectively; , The revenue from material recycling for batteries and supercapacitors, respectively; , These represent the remaining lifespan percentages of batteries and supercapacitors, respectively. , These are the residual coefficients for batteries and supercapacitors, respectively. , These are the material recycling coefficients for batteries and supercapacitors, respectively.

[0079] The main constraints considered in the life-cycle cost model are as follows:

[0080] Energy constraints:

[0081]

[0082]

[0083] In the formula: This represents the lower limit of the remaining battery capacity. Let k be the remaining battery capacity at time k. This represents the maximum remaining battery capacity. This represents the lower limit of the remaining capacity of the supercapacitor. The remaining capacitance of the supercapacitor at time k; This represents the upper limit of the remaining capacity of the supercapacitor.

[0084] E / P constraint: The continuous operating time of the supercapacitor at rated power is set to 10 min to 60 min;

[0085]

[0086]

[0087] As a preferred embodiment of the dual-layer hybrid energy storage capacity configuration method for a wind-storage combined power generation system described in this invention, the specific steps in step 3 are as follows: This capacity configuration method quantifies and evaluates battery lifespan based on rainflow counting and an equivalent cycle life model to obtain battery lifespan, and finally calculates the total lifespan cost. This cost function is then integrated into a particle swarm optimization module. The algorithm continuously searches for system parameter configurations that minimize the total lifespan cost by initializing particles and iteratively updating individual and swarm optimal solutions. The capacity configuration optimization steps for the hybrid energy storage system based on the total lifespan cost are as follows:

[0088] (1) Input the battery power curve obtained by MPC decomposition, calculate the battery's state of charge and operating parameters, and extract the charge-discharge cycle and its discharge depth in operation based on the rainflow counting method.

[0089] (2) Based on the extracted depth of discharge, the equivalent cycle life method is used to convert the cycle under complex actual working conditions into the standard equivalent cycle number, thereby accurately calculating the degree of battery life decay under the working conditions and obtaining the battery's operating life.

[0090] (3) Construct a life-cycle cost objective function and establish a capacity configuration model for hybrid energy storage systems;

[0091] (4) The particle swarm optimization algorithm is used to solve the capacity configuration model. The parameters of the particle swarm optimization algorithm are initialized, and the particle positions are updated iteratively to continuously search for the optimal configuration that minimizes the total life cycle cost.

[0092] (5) After the optimization algorithm reaches the preset number of iterations, it outputs the global optimal solution.

[0093] As a preferred embodiment of the dual-layer hybrid energy storage capacity configuration method for a wind-storage combined power generation system described in this invention, the specific steps of dynamically adjusting the MPC weight parameters based on the configuration results in step 3 are as follows: 1) Dynamic weight adjustment based on Sigmoid function mapping:

[0094] Obtain the total life-cycle cost of batteries and supercapacitors to determine the cost percentage of the energy storage device:

[0095]

[0096]

[0097] In the formula: This represents the proportion of batteries in the total energy storage investment. This represents the proportion of supercapacitors in total energy storage investment; Cost over the entire battery lifecycle; Cost over the entire lifespan of a supercapacitor;

[0098] The Sigmoid function is used to smoothly map the battery cost ratio to the [0, 1] interval, avoiding abrupt changes and allowing the feedback factor to change continuously. The Sigmoid function is shown below:

[0099]

[0100] In the formula: As a feedback factor; The center point of the mapping is k; k is the slope parameter.

[0101] Based on the feedback factor, the adjusted weighting coefficients for the battery and supercapacitor are calculated separately:

[0102]

[0103]

[0104] In the formula: As the attenuation factor, when When the decay rate drops to less than 0.3, its value becomes 0.3;

[0105] The attenuation factor is expressed as follows:

[0106]

[0107] In the formula: This represents the current iteration number;

[0108] 2) Weight adjustment based on lifetime feedback:

[0109] Battery life feedback: A life feedback factor is generated based on the margin size to adjust the battery usage weight. The life margin is calculated as follows:

[0110]

[0111] Where: M is the battery life margin; To meet the minimum cycle life requirement, this paper sets it to 5 years;

[0112] A piecewise linear rule is adopted to form a lifetime feedback factor based on the lifetime margin;

[0113] Based on the obtained lifetime feedback factor and feedback strength, the following adjustments and updates are made:

[0114]

[0115] In the formula: For adjustment factors; Lifetime feedback factor; For feedback strength;

[0116] Battery usage weight updated to: .

[0117] Compared with existing technologies, the advantages of this invention are as follows: The inner layer employs MPC (Multi-Level Configuration) to dynamically and optimally decompose the given total power demand of hybrid energy storage into real-time output commands from batteries and supercapacitors. The outer layer, based on the battery power sequence output by MPC, uses rainflow counting and equivalent cycle life methods to quantify battery life loss, establishing a refined cost model for the entire life cycle. Furthermore, it utilizes particle swarm optimization to globally optimize the rated capacity and power of the energy storage system. Finally, feedback information is generated based on the obtained capacity configuration results, dynamically adjusting the MPC weight parameters to form a closed-loop feedback mechanism. This method effectively reduces energy storage configuration costs by 31.62% and is suitable for energy storage optimization in large-scale wind power grid connection applications. Attached Figure Description

[0118] To more clearly illustrate the technical solutions of the embodiments of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and detailed embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0119] Figure 1 This is a schematic diagram of the rainflow counting method of the present invention, wherein (a) is the original SOC curve; (b) is the extreme SOC curve; (c) is the cycle 1 diagram; and (d) are the cycle 2 and 3 diagrams.

[0120] Figure 2 The diagram shows the relationship between cycle life and depth of discharge in this invention, where (a) is the relationship between battery depth of discharge and cycle life; and (b) is the residual analysis diagram.

[0121] Figure 3 This is a diagram illustrating the capacity configuration method for a hybrid energy storage system based on total life-cycle cost according to the present invention.

[0122] Figure 4 This is a target power diagram for the hybrid energy storage of the present invention;

[0123] Figure 5 This is a diagram showing the target power allocation results of the hybrid energy storage based on VMD+GWO according to the present invention;

[0124] Figure 6 This is a diagram showing the target power allocation results of the hybrid energy storage based on MPC according to the present invention. Detailed Implementation

[0125] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0126] Secondly, the present invention is described in detail with reference to the schematic diagrams. When detailing the embodiments of the present invention, for ease of explanation, the cross-sectional views illustrating the device structure may be partially enlarged, not according to the usual scale. Furthermore, the schematic diagrams are merely examples and should not limit the scope of protection of the present invention. In addition, actual fabrication should include three-dimensional spatial dimensions of length, width, and depth.

[0127] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0128] This invention provides a method for configuring dual-layer hybrid energy storage capacity in a wind-storage combined power generation system. By introducing factors such as rigid body external distance, visual geometric saliency factor, and dynamic motion intention, the traditional physical risk field is modified, thereby constructing a risk distribution model that conforms to the driver's personalized psychological expectations.

[0129] Specifically, a method for configuring a two-layer hybrid energy storage capacity for a wind-storage combined power generation system includes the following steps:

[0130] Step 1: Decompose the target power of hybrid energy storage using a model predictive control algorithm;

[0131] Step 1.1: Establish the MPC weighted quadratic objective function;

[0132] To ensure that the total output of the energy storage system strictly follows the target power of the hybrid energy storage, the objective function first considers power point tracking accuracy. Secondly, based on equipment cost characteristics, power allocation is guided by differentiated weights—supercapacitors are encouraged to handle high-frequency fluctuations due to their low power cost, while batteries are discouraged due to their high power cost to reduce investment. Simultaneously, the objective function also incorporates a power smoothing term to suppress drastic power fluctuations and extend equipment lifespan. In summary, the objective function is set as a multi-objective weighted quadratic function:

[0133]

[0134] Where: N p For prediction in the time domain; Let k be the power that the lithium-ion battery needs to consume at time k; Let k be the power that the supercapacitor needs to absorb at time k. Let Q1 be the target power of hybrid energy storage at time k; Q2 be the weight of the power tracking accuracy of target 1; Q3 be the weight of the battery usage cost of target 2; Q4 be the weight of the supercapacitor usage cost of target 3; and Q5 be the weight of the power smoothing term of target 4. Let k be the change in the power required to be consumed by the lithium-ion battery at time k compared to time k-1. This represents the change in the power that the supercapacitor needs to absorb at time k compared to time k-1.

[0135] Step 1.2: Construct a power command allocation model based on MPC to achieve effective decomposition of target power for hybrid energy storage;

[0136] When MPC is applied to the power decomposition of hybrid energy storage systems, its core is to construct a dynamic model of the critical states and power distribution relationships of the energy storage elements. The discrete-space state equation of the system is as follows:

[0137]

[0138] In the formula: Let A be the predicted value of the state variable at time k+1; A is the state matrix, describing how the state variable transitions from time k to time k+1. Let B1 be the state variable at time k; B1 is the input matrix, describing how the control variables affect the changes in the state variables. B1 represents the control variable at time k; B2 is the disturbance matrix, describing the influence of the disturbance variable on the state variable. Let k be the perturbation variable at time k;

[0139] Further derivation yields the predicted state variables at time k+2:

[0140]

[0141] Since the target power of the hybrid energy storage system is a known command given by the upper-level energy management system, it is not a "disturbance" but a "target" that the controller needs to track. Therefore, it is written into the equality constraints, and the disturbance variable is ignored. The space state equation simplifies to:

[0142]

[0143] Based on the equality constraints in the power allocation process of the hybrid energy storage system, the power allocation space state model based on the MPC algorithm can be obtained:

[0144]

[0145] In the formula: This represents the state of charge of the lithium-ion battery at time k+1. This represents the state of charge of the lithium-ion battery at time k. This represents the state of charge of the supercapacitor at time k+1. This represents the state of charge of the supercapacitor at time k; , The average efficiency of batteries and supercapacitors; , Rated capacity for a battery and supercapacitor configuration;

[0146] Will and As a system state variable:

[0147]

[0148] Will and As a control variable of the system:

[0149]

[0150] Combining the two, we obtain the complete discrete state-space equation:

[0151]

[0152] The main constraints to consider in the power command allocation model are as follows:

[0153] (1) Equality constraints

[0154]

[0155] (2) SOC constraint

[0156] To prevent overcharging and over-discharging of energy storage devices during wind power fluctuation smoothing, the state of charge of energy storage should comply with the following constraints:

[0157]

[0158]

[0159] In the formula: The battery is in its state of charge. This is the state of charge of the supercapacitor.

[0160] (3) Power constraints

[0161]

[0162]

[0163] In the formula: This refers to the lower limit of battery charging and discharging power. This refers to the upper limit of battery charging and discharging power. This refers to the lower limit of the charging and discharging power of supercapacitors. This represents the upper limit of the charging and discharging power of the supercapacitor.

[0164] Step 2: Use the rainflow counting method and the equivalent cycle life method to quantify battery life loss and establish a refined cost model for the entire life cycle;

[0165] Step 2.1: Calculate the depth of discharge using the rainflow counting method;

[0166] Estimating battery life based on a battery capacity decay model requires obtaining the depth of discharge (DOD) for each cycle. However, due to the irregular operation of batteries under different application scenarios, it is difficult to determine the DOD and cycle number from the charge-discharge curve. Rainflow counting can decompose and simplify complex, randomly fluctuating load-time histories (such as changes in battery charging and discharging current and voltage) into a series of complete "cycles" for fatigue or life prediction. This paper uses a simplified rainflow counting method to obtain the DOD of the battery in each operation. The principle is as follows: Figure 1 As shown:

[0167] (1) Obtain the sequence of battery state of charge changes over time. Figure 1 As shown in (a), the maximum and minimum values ​​in the sequence are found, and small fluctuations such as measurement noise or irrelevant fluctuations are filtered out. The extreme values ​​are then connected sequentially. Figure 1 As shown in (b);

[0168] (2) Rotate the processed SOC curve 90° clockwise, resembling a series of roof shapes. Rainwater flows down the roof successively at the starting point and at each extreme point;

[0169] (3) The rain flows to the peak point (i.e., the eaves) and drips vertically. If there is another eave below the eaves, the rain drips to the lower eaves and continues to flow downwards, as in the process B→B'; if there is no obstruction below, the rain changes direction and flows upwards, as in the process D→E.

[0170] (4) When the rainwater flow encounters raindrops flowing down from the roof above, it stops flowing and forms a cycle;

[0171] (5) Draw each cycle according to the starting point and ending point of the rain flow process, extract all cycles in sequence, and record their peak and valley values. The peak and valley difference of each rain flow is the discharge depth of the cycle.

[0172] Figure 1 As shown in (c), this is cycle 1. The rainwater flows from point A to point B, then along the roof eaves CD. At point D, it changes direction and flows along the roof eaves DE to point E, where it falls to the roof eaves FG, continuing until it reaches point G, ending the cycle. This equivalent cycle is A→D→G, with a discharge depth of 0.8-0.28=0.52 and an average SOC level of (0.8+0.28) / 2=0.54. Besides this cycle, there are also cycles 2 and 3. Figure 1 As shown in (d), cycle 2 is B→C→B', with a discharge depth of 0.7-0.4=0.3 and an average SOC level of (0.7+0.4) / 2=0.55; cycle 3 is E→F→E', with a discharge depth of 0.6-0.4=0.2 and an average SOC level of (0.6+0.4) / 2=0.5. The battery life can be obtained by substituting the cycle parameters obtained from the rainflow counting method into the battery capacity decay model.

[0173] Step 2.2: Calculate battery life using the equivalent cycle life method;

[0174] In wind power fluctuation mitigation scenarios, to ensure the safety of energy storage operation, battery temperature is typically controlled within a certain range. Therefore, to simplify calculations, only the impact of DOD level on battery capacity degradation is considered. Table 1 shows the cycle life of a certain battery model at different depths of discharge:

[0175] Table 1: Relationship between battery depth of discharge and cycle life

[0176]

[0177] Based on the data in Table 1, an exponential decay curve was used for fitting, resulting in the fitted curve. Figure 2 As shown in (a), the obtained double exponential decay function is used to characterize the relationship between the depth of discharge and cycle life:

[0178]

[0179] In the formula: This refers to the number of battery cycles. This refers to the depth of charge and discharge of the battery.

[0180] Perform residual analysis on the fitting process Figure 2 As shown in (b), the mean residual of the curve fitting is 0.17, which is very close to zero, and the standard deviation of the residual is 160.89. Statistical indicators show that the model fit is relatively accurate.

[0181] In actual charging and discharging processes, the depth of discharge of a battery changes in real time. Using the equivalent cycle life method, the charge-discharge cycles with different depths of discharge experienced by the battery in actual operation are converted into the standard number of cycles at 100% depth of discharge, based on the principle of equivalent life loss. The expression for the equivalent cycle number is then obtained through calculation:

[0182]

[0183] In the formula: This represents the equivalent number of iterations. The number of cycles; This represents the number of cycles the battery corresponds to when the depth of discharge is 1. For the battery at a depth of discharge of The number of loops corresponding to the time.

[0184]

[0185] In the formula: T represents the battery life loss. When T=1 or = At that time, the battery's lifespan will be exhausted.

[0186] Step 2.3: Establish a capacity configuration model for hybrid energy storage systems based on total life cycle cost;

[0187] Based on the equivalent cycle life of the battery obtained in the previous section, a refined life-cycle cost model for an energy storage system is constructed. This model takes minimizing the total life-cycle cost as its objective function and comprehensively covers initial investment, auxiliary equipment, long-term operation and maintenance, equipment replacement, and final disposal and recycling revenue.

[0188]

[0189] In the formula: For investment and installation costs; To cover auxiliary costs; For operation and maintenance costs; For replacement costs; For the revenue from scrapping and recycling.

[0190] (1) Investment and installation costs:

[0191]

[0192]

[0193]

[0194]

[0195]

[0196] In the formula: , These are the investment costs for batteries and supercapacitors, respectively. , The installation costs are for batteries and supercapacitors, respectively. , These are the unit power configuration costs for batteries and supercapacitors, respectively. , These are the unit capacity configuration costs for batteries and supercapacitors, respectively. , The rated power is configured for the battery and the supercapacitor, respectively; , These are the rated capacities for the battery and supercapacitor configurations, respectively. , These are the installation coefficients for batteries and supercapacitors, respectively.

[0197] (2) Ancillary costs:

[0198]

[0199]

[0200]

[0201] In the formula: , These are the auxiliary costs for batteries and supercapacitors, respectively. , These are the cost coefficients for auxiliary equipment for batteries and supercapacitors, respectively.

[0202] (3) Operation and maintenance costs:

[0203]

[0204]

[0205]

[0206]

[0207] In the formula: , These are the operating and maintenance costs of batteries and supercapacitors, respectively. , These are the operation and maintenance cost coefficients for batteries and supercapacitors, respectively; PWF is the present value factor of an equal annuity; T is the project analysis period. The benchmark discount rate is used.

[0208] (4) Replacement cost:

[0209]

[0210]

[0211] In the formula: n is the number of battery replacements during the project cycle; denoted as , where i represents the battery's lifespan; i represents the i-th replacement, with replacement cost taking into account cost reductions due to technological advancements (3% annual decrease).

[0212] (5) Revenue from scrapping and recycling:

[0213]

[0214]

[0215]

[0216]

[0217]

[0218] In the formula: , The residual value gains for batteries and supercapacitors, respectively; , The revenue from material recycling for batteries and supercapacitors, respectively; , These represent the remaining lifespan percentages of batteries and supercapacitors, respectively. , These are the residual coefficients for batteries and supercapacitors, respectively. , These are the material recycling coefficients for batteries and supercapacitors, respectively.

[0219] The main constraints considered in the life-cycle cost model are as follows:

[0220] (1) Energy constraint:

[0221]

[0222]

[0223] In the formula: This represents the lower limit of the remaining battery capacity. Let k be the remaining battery capacity at time k. This represents the maximum remaining battery capacity. This represents the lower limit of the remaining capacity of the supercapacitor. The remaining capacitance of the supercapacitor at time k; This represents the upper limit of the remaining capacity of the supercapacitor.

[0224] (2) E / P constraint:

[0225] The battery can efficiently handle low-frequency, smooth power fluctuations ranging from minutes to hours. To ensure the smoothing effect while avoiding cost waste caused by over-configuration of capacity, the battery's continuous operating time at rated power is set to 30-120 minutes. The supercapacitor has an extremely fast response and ultra-long cycle life, and can efficiently handle instantaneous high-frequency fluctuations. Therefore, the supercapacitor's continuous operating time at rated power is set to 10-60 minutes.

[0226]

[0227]

[0228] Step 3: Solve the lifecycle cost model using the particle swarm optimization algorithm, and adjust the MPC weights based on the solution results.

[0229] Step 3.1: Solve the hybrid energy storage capacity configuration model using the particle swarm optimization algorithm;

[0230] This capacity configuration method uses rainflow counting and an equivalent cycle life model to quantitatively assess battery life loss, obtain battery life, and ultimately calculate the total lifespan cost. This cost function is integrated into a particle swarm optimization module. The algorithm initializes particles, iteratively updates individual and swarm optimal solutions, and continuously searches for system parameter configurations that minimize the total lifespan cost. The capacity configuration optimization process for hybrid energy storage systems based on total lifespan cost is as follows: Figure 3 As shown, the steps are as follows:

[0231] (1) Input the battery power curve obtained by MPC decomposition, calculate the battery's state of charge and operating parameters, and extract the charge-discharge cycle and its discharge depth in operation based on the rainflow counting method.

[0232] (2) Based on the extracted depth of discharge, the equivalent cycle life method is used to convert the cycle under complex actual working conditions into the standard equivalent cycle number, thereby accurately calculating the degree of battery life decay under the working conditions and obtaining the battery's operating life.

[0233] (3) Construct a life-cycle cost objective function and establish a capacity configuration model for hybrid energy storage systems;

[0234] (4) The capacity configuration model is solved using the particle swarm optimization algorithm. The parameters of the particle swarm optimization algorithm are initialized, and the particle positions are updated iteratively to continuously search for the optimal configuration that minimizes the total life cycle cost;

[0235] (5) After the optimization algorithm reaches the preset number of iterations, it outputs the global optimal solution.

[0236] Step 3.2: Dynamically adjust MPC weight parameters based on configuration results;

[0237] (1) Dynamic weight adjustment based on Sigmoid function mapping:

[0238] Based on the results of the previous round of optimization, the feedback module can obtain the total life cycle cost of the battery and supercapacitor, and thus obtain the cost ratio of the energy storage device:

[0239]

[0240]

[0241] In the formula: This represents the proportion of batteries in the total energy storage investment. This represents the proportion of supercapacitors in total energy storage investment; Cost over the entire battery lifecycle; This refers to the total lifecycle cost of a supercapacitor.

[0242] Subsequently, the Sigmoid function is used to smoothly map the battery cost ratio to the [0, 1] interval, avoiding abrupt changes and allowing the feedback factor to change continuously. The Sigmoid function is shown below:

[0243]

[0244] In the formula: As a feedback factor; The center point of the mapping is set to 0.5 to symmetrically handle the cost ratio; k is the slope parameter, which controls the steepness of the curve, and is set to 10 in this paper.

[0245] The Sigmoid function plays a crucial role in dynamic weight adjustment mechanisms. Its purpose is to non-linearly convert cost proportions into feedback factors for adjusting weights, making weight changes both smooth and discriminative.

[0246] When x deviates by 50%, the weight adjustment responds quickly. For example, when x=0.4 changes to x=0.6, although the cost ratio only changes by 0.2, the feedback factor jumps from 0.27 to 0.73. The magnitude of the weight adjustment changes significantly, effectively guiding the optimization direction.

[0247] When x approaches 1, the feedback factor approaches 1. The weight of battery usage increases, while the weight of supercapacitors decreases.

[0248] When x approaches 0, the feedback factor approaches 0. The weight of battery usage decreases, while the weight of supercapacitors increases.

[0249] The flat regions at both ends of the Sigmoid function represent a soft saturation characteristic, meaning that when the battery cost ratio is in an extreme case, the system's sensitivity to cost changes decreases and the weight adjustment tends to stabilize.

[0250] Based on the feedback factor, the adjusted weighting coefficients for the battery and supercapacitor are calculated separately:

[0251]

[0252]

[0253] In the formula: As the attenuation factor, when When the decay rate drops to less than 0.3, its value is 0.3.

[0254] In dynamic weight adjustment, the decay factor is a key parameter that controls the gradual decrease of the adjustment amplitude with the number of iterations. Its role is to allow for larger weight adjustments in the early stages of optimization to quickly explore favorable cost allocation directions; while in the later stages of optimization, reducing the adjustment amplitude avoids oscillations caused by drastic weight changes, ensuring the convergence stability of the algorithm. Simultaneously, by setting a lower limit (0.3 in this paper), a certain degree of adaptability is retained, allowing the system to still respond to cost changes in the later stages of iteration and maintain the ability to fine-tune the optimization direction. The decay factor is expressed as follows:

[0255]

[0256] In the formula: This represents the current iteration number.

[0257] (2) Weight adjustment based on lifetime feedback:

[0258] Battery life feedback: The battery's cycle life should exceed the minimum life requirement set in this paper, with a certain margin. Therefore, this paper generates a life feedback factor based on the margin to adjust the battery usage weight. The life margin calculation is as follows:

[0259]

[0260] Where: M is the battery life margin; To meet the minimum cycle life requirement, this paper sets it to 5 years.

[0261] A piecewise linear rule is adopted to form a lifetime feedback factor based on the lifetime margin, as shown in Table 2.

[0262] Table 2: Lifetime Protection Feedback Factor Setting Table Based on Operating Margin

[0263]

[0264] The feedback intensity is dynamically adjusted during the feedback process to control the impact of feedback information on the MPC weights in each iteration. Based on the cost improvement of the current iteration, the intensity of the feedback in the next round is adaptively adjusted, thereby balancing exploration and convergence during the optimization process and avoiding oscillations or divergences caused by over-adjustment.

[0265] If the configuration cost increases after iterative feedback, it indicates that the feedback has been over-adjusted and performance has degraded. The feedback strength should be significantly reduced (set to 0.3) to stabilize the optimization process. If the cost improves but the improvement is small, it indicates that the solution is close to the optimal solution, but it still needs to be adjusted carefully. The feedback strength should be moderately reduced (set to 0.7) to avoid small oscillations. If the configuration cost improves significantly, it indicates that the current feedback direction is correct. The normal strength should be maintained (set to 1.0) and the feedback information should continue to be fully utilized to guide the optimization.

[0266] Based on the obtained lifetime feedback factor and feedback strength, the following adjustments and updates are made:

[0267]

[0268] In the formula: For adjustment factors; Lifetime feedback factor; For feedback strength;

[0269] Battery usage weight updated to:

[0270]

[0271] Step 3.3: Hybrid energy storage system capacity configuration;

[0272] The relevant parameters for the capacity configuration of the energy storage system are shown in Table 3.

[0273] Table 3: Technical and Economic Parameters Related to Energy Storage Devices

[0274]

[0275] To compare the advantages and disadvantages of different methods, the following two schemes are set up for comparative verification:

[0276] Option 1: Use GWO-VMD to decompose the target power of hybrid energy storage and configure capacity based on the life-cycle cost model. This option predetermines the power that the energy storage unit needs to absorb through fixed frequency domain decomposition and reconstruction, and optimizes capacity configuration under this scenario. However, there is no coordinated optimization between operation control and capacity configuration.

[0277] Option 2: Use MPC to decompose the target power of hybrid energy storage, configure capacity based on the life cycle cost model, and have a feedback iterative process.

[0278] The target power of a hybrid energy storage system is the difference between the original wind power and the grid-connected power. Essentially, it represents the total target power demand that the hybrid energy storage system needs to absorb or release the power component that requires its consumption. Figure 4 As shown.

[0279] Based on the standard quadratic programming form of MPC, the solution is obtained by calling MATLAB's built-in quadprog solver, yielding power control commands for the battery and supercapacitor. The two work together to achieve precise tracking of the target power of the hybrid energy storage system. The final power to be absorbed by the supercapacitor and battery is as follows: Figure 5 and Figure 6 As shown.

[0280] Based on the power that the energy storage device needs to absorb, obtained from MPC decomposition, a refined life-cycle cost model is further established. Using the minimum system life-cycle cost as the objective function, a particle swarm optimization algorithm is employed to optimize the energy storage capacity configuration. The resulting optimal configuration minimizes the life-cycle cost while meeting grid connection fluctuation constraints. Subsequently, the optimized cost result is transformed into a feedback factor, dynamically adjusting the weight coefficients of the objective function in the MPC controller, enabling the control strategy to balance economy and tracking performance during operation.

[0281] After optimization using the particle swarm optimization algorithm, the optimal capacity configuration results of the hybrid energy storage system are shown in Table 4. Table 4 shows that Method 2 has lower investment and total costs than Method 1. Compared to Method 1, Method 2 reduces costs by 20.8835 million yuan, proving that the proposed capacity configuration method for a two-layer hybrid energy storage system based on MPC feedback and life-cycle cost optimization is superior to the VMD process of first performing frequency decomposition and then capacity configuration. This advantage mainly stems from the closed-loop feedback mechanism adopted by Method 2, which couples MPC power allocation with PSO capacity optimization iteration, enabling the power allocation strategy to dynamically adapt to changes in capacity configuration. In contrast, the separation of frequency decomposition and capacity configuration in Method 1 fails to consider the mutual influence between power allocation and capacity, resulting in poor economic efficiency in the final configuration.

[0282] Table 4: Capacity Configuration Results

[0283]

[0284] Although the present invention has been described above with reference to embodiments, various modifications can be made and components can be replaced with equivalents without departing from the scope of the invention. In particular, as long as there is no structural conflict, the features in the disclosed embodiments can be combined with each other in any manner. The lack of an exhaustive description of these combinations in this specification is merely for the sake of brevity and resource conservation. Therefore, the present invention is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A method for configuring a two-layer hybrid energy storage capacity in a wind-storage combined power generation system, characterized in that, Includes the following steps: Step 1: Decompose the target power of hybrid energy storage using MPC: Establish an MPC weighted quadratic objective function, construct a power command allocation model based on MPC, and realize the effective decomposition of the target power of hybrid energy storage; Step 2: Calculate the depth of discharge using the rainflow counting method and the battery life using the equivalent cycle life method, and establish a capacity configuration model for the hybrid energy storage system based on the total life cycle cost. Step 3: Use the particle swarm optimization algorithm to solve the hybrid energy storage capacity configuration model, and dynamically adjust the MPC weight parameters based on the configuration results.

2. The method for configuring a dual-layer hybrid energy storage capacity for a wind-storage combined power generation system according to claim 1, characterized in that, The specific steps for establishing the MPC weighted quadratic objective function in step 1 are as follows: the objective function is set as a multi-objective weighted quadratic function: Where: N p For prediction in the time domain; Let k be the power that the lithium-ion battery needs to consume at time k; Let k be the power that the supercapacitor needs to absorb at time k. Let Q1 be the target power of hybrid energy storage at time k; Q2 be the weight of the power tracking accuracy of target 1; Q3 be the weight of the battery usage cost of target 2; Q4 be the weight of the supercapacitor usage cost of target 3; and Q5 be the weight of the power smoothing term of target 4. Let k be the change in the power required to be consumed by the lithium-ion battery at time k compared to time k-1. This represents the change in the power that the supercapacitor needs to absorb at time k compared to time k-1.

3. The method for configuring a dual-layer hybrid energy storage capacity for a wind-storage combined power generation system according to claim 1, characterized in that, The specific steps for constructing the MPC-based power command allocation model in step 1 are as follows: The discrete-space state equation of the system is as follows: In the formula: Let A be the predicted value of the state variable at time k+1; A is the state matrix, describing how the state variable transitions from time k to time k+1. Let B1 be the state variable at time k; B1 is the input matrix, describing how the control variables affect the changes in the state variables. B1 represents the control variable at time k; B2 is the disturbance matrix, describing the influence of the disturbance variable on the state variable. Let k be the perturbation variable at time k; Further derivation yields the predicted state variables at time k+2: The space state equation simplifies to: Based on the equality constraints in the power allocation process of the hybrid energy storage system, a power allocation space state model based on the MPC algorithm is obtained: In the formula: This represents the state of charge of the lithium-ion battery at time k+1. This represents the state of charge of the lithium-ion battery at time k. This represents the state of charge of the supercapacitor at time k+1. This represents the state of charge of the supercapacitor at time k; , The average efficiency of batteries and supercapacitors; , Rated capacity for batteries and supercapacitors; Will and As a system state variable: Will and As a control variable of the system: Combining the two, we obtain the complete discrete state-space equation: 。 4. The method for configuring a dual-layer hybrid energy storage capacity for a wind-storage combined power generation system according to claim 3, characterized in that, The main constraints to consider in the power command allocation model are as follows: Equality constraints: SOC Constraints: To prevent overcharging and over-discharging of energy storage devices during wind power fluctuation smoothing, the state of charge (SOC) of energy storage should comply with the following constraints: In the formula: The battery is in its state of charge. This is the state of charge of a supercapacitor; Power constraints: In the formula: This refers to the lower limit of battery charging and discharging power. This refers to the upper limit of battery charging and discharging power. This refers to the lower limit of the charging and discharging power of supercapacitors. This represents the upper limit of the charging and discharging power of the supercapacitor.

5. A method for configuring dual-layer hybrid energy storage capacity in a wind-storage combined power generation system according to claim 1, characterized in that, The specific steps for calculating the depth of discharge using the rainflow counting method in step 2 are as follows: A simplified rainflow counting method is used to obtain the DOD of the battery in each operation. The cycle parameters obtained by the rainflow counting method are then substituted into the battery capacity decay model to obtain the battery life.

6. The method for configuring a dual-layer hybrid energy storage capacity for a wind-storage combined power generation system according to claim 1, characterized in that, The specific steps for calculating battery life using the equivalent cycle life method in step 2 are as follows: only considering the impact of DOD level on battery capacity decay, an exponential decay function is used for fitting, and the resulting double exponential decay function is used to characterize the relationship between depth of discharge and cycle life. In the formula: This refers to the number of battery cycles. The depth of charge and discharge of the battery; The equivalent cycle life method is used to convert the charge-discharge cycles with different depths of discharge experienced by the battery in actual operation into standard cycle numbers at 100% depth of discharge, according to the principle of equivalent life loss. The expression for the equivalent cycle number is then obtained through calculation: In the formula: This represents the equivalent number of iterations. The number of cycles; This represents the number of cycles the battery corresponds to when the depth of discharge is 1. For the battery at a depth of discharge of The number of loops corresponding to the time; In the formula: T represents the battery life loss.

7. A method for configuring a dual-layer hybrid energy storage capacity for a wind-storage combined power generation system according to claim 1, characterized in that, The specific steps for establishing a hybrid energy storage system capacity configuration model based on total life cycle cost in step 2 are as follows: The objective function is to minimize total life cycle cost, comprehensively covering initial investment, auxiliary equipment, long-term operation and maintenance, equipment replacement, and final disposal and recycling revenue. In the formula: For investment and installation costs; To cover auxiliary costs; For operation and maintenance costs; For replacement costs; For the revenue from scrap recycling; Investment and installation costs: In the formula: , These are the investment costs for batteries and supercapacitors, respectively. , The installation costs are for batteries and supercapacitors, respectively. , These are the unit power configuration costs for batteries and supercapacitors, respectively. , These are the unit capacity configuration costs for batteries and supercapacitors, respectively. , The rated power is configured for the battery and the supercapacitor, respectively; , These are the rated capacities for the battery and supercapacitor configurations, respectively. , These are the installation coefficients for batteries and supercapacitors, respectively. Ancillary costs: In the formula: , These are the auxiliary costs for batteries and supercapacitors, respectively. , These are the cost coefficients for auxiliary equipment for batteries and supercapacitors, respectively. Operation and maintenance costs: In the formula: , These are the operating and maintenance costs of batteries and supercapacitors, respectively. , These are the operation and maintenance cost coefficients for batteries and supercapacitors, respectively; PWF is the present value factor of an equal annuity; T is the project analysis period. The benchmark discount rate; Replacement cost: In the formula: n is the number of battery replacements during the project cycle; denoted as , where is the battery's lifespan; represents the i-th replacement, with replacement cost taking into account cost reductions due to technological advancements. Revenue from scrap recycling: In the formula: , The residual value gains for batteries and supercapacitors, respectively; , The revenue from material recycling for batteries and supercapacitors, respectively; , These represent the remaining lifespan percentages of batteries and supercapacitors, respectively. , These are the residual coefficients for batteries and supercapacitors, respectively. , These are the material recycling coefficients for batteries and supercapacitors, respectively. The main constraints considered in the life-cycle cost model are as follows: Energy constraints: In the formula: This represents the lower limit of the remaining battery capacity. Let K be the remaining battery capacity at time k. This represents the maximum remaining battery capacity. This represents the lower limit of the remaining capacity of the supercapacitor. The remaining capacitance of the supercapacitor at time k; This represents the upper limit of the remaining capacity of the supercapacitor. E / P constraint: The continuous operating time of the supercapacitor at rated power is set to 10 min to 60 min; 。 8. A method for configuring dual-layer hybrid energy storage capacity in a wind-storage combined power generation system according to claim 1, characterized in that, The specific steps in step 3 are as follows: The capacity configuration method uses the rainflow counting method and the equivalent cycle life model to quantitatively evaluate the battery's lifespan, obtain the battery life, and finally calculate the total lifespan cost. This cost function is connected to the particle swarm optimization module. The algorithm initializes particles, iteratively updates the optimal solutions for individuals and the swarm, and continuously searches for system parameter configurations that minimize the total lifespan cost. The capacity configuration optimization steps for the hybrid energy storage system based on the total lifespan cost are as follows: (1) Input the battery power curve obtained by MPC decomposition, calculate the battery's state of charge and operating parameters, and extract the charge-discharge cycle and its discharge depth in operation based on the rainflow counting method. (2) Based on the extracted depth of discharge, the equivalent cycle life method is used to convert the cycle under complex actual working conditions into the standard equivalent cycle number, thereby accurately calculating the degree of battery life decay under the working conditions and obtaining the battery's operating life. (3) Construct a life-cycle cost objective function and establish a capacity configuration model for hybrid energy storage systems; (4) The particle swarm optimization algorithm is used to solve the capacity configuration model. The parameters of the particle swarm optimization algorithm are initialized, and the particle positions are updated iteratively to continuously search for the optimal configuration that minimizes the total life cycle cost. (5) After the optimization algorithm reaches the preset number of iterations, it outputs the global optimal solution.

9. A method for configuring a dual-layer hybrid energy storage capacity for a wind-storage combined power generation system according to claim 1, characterized in that, The specific steps for dynamically adjusting the MPC weight parameters based on the configuration results in step 3 are as follows: 1) Dynamic weight adjustment based on Sigmoid function mapping: Obtain the total life-cycle cost of batteries and supercapacitors to determine the cost percentage of the energy storage device: In the formula: This represents the proportion of batteries in the total energy storage investment. This represents the proportion of supercapacitors in total energy storage investment; Cost over the entire battery lifecycle; Cost over the entire lifespan of a supercapacitor; The Sigmoid function is used to smoothly map the battery cost ratio to the [0, 1] interval, avoiding abrupt changes and allowing the feedback factor to change continuously. The Sigmoid function is shown below: In the formula: As a feedback factor; The center point of the mapping is k; k is the slope parameter. Based on the feedback factor, the adjusted weighting coefficients for the battery and supercapacitor are calculated separately: In the formula: As the attenuation factor, when When the decay rate drops to less than 0.3, its value becomes 0.3; The attenuation factor is expressed as follows: In the formula: This represents the current iteration number; 2) Weight adjustment based on lifetime feedback: Battery life feedback: A life feedback factor is generated based on the margin size to adjust the battery usage weight. The life margin is calculated as follows: Where: M is the battery life margin; To meet the minimum cycle life requirement, this paper sets it to 5 years; A piecewise linear rule is adopted to form a lifetime feedback factor based on the lifetime margin; Based on the obtained lifetime feedback factor and feedback strength, the following adjustments and updates are made: In the formula: For adjustment factors; Lifetime feedback factor; For feedback strength; Battery usage weight updated to: .

Images