A parameter optimization method for adaptive damping inertia adjustment of a network-constructed optical storage system

Abstract

This invention discloses a parameter optimization method for adaptive damped inertia adjustment in a grid-connected photovoltaic-storage system, belonging to the field of new energy power generation control and smart grid technology. The invention establishes a second-order nonlinear coupled dynamic model of photovoltaic-storage-grid, constructs a multi-objective optimization function that considers system stability, equipment losses, and coupling suppression, and uses an improved particle swarm optimization algorithm incorporating coupled dynamic feedback to solve for the optimal parameters. A closed-loop adaptive adjustment mechanism of "state perception - parameter decision - smooth update - feedback correction" is then established, and the method is verified through simulation and testing. This method effectively suppresses multi-device coupled oscillations, improves the system's voltage-frequency coordinated stability and dynamic response robustness, and provides reliable technical support for high-proportion grid-connected new energy systems.

Description

Technical Field

[0001] This invention relates to the field of new energy power generation control and smart grid technology, and in particular to a parameter optimization method for adaptive damping inertia adjustment of a grid-type photovoltaic-storage system. Background Technology

[0002] The grid-connection penetration rate of new energy power generation systems centered on grid-connected photovoltaic (PV) and energy storage (ESS) continues to climb. Grid-connected PV-ESN systems, by simulating the inertia and damping characteristics of synchronous generators, provide crucial voltage-frequency support to the power grid and are one of the core technologies for ensuring the safe and stable operation of a grid with a high proportion of new energy connected to the grid. However, with the increase in the number of PV and ESN devices and the expansion of grid-connected scale, a complex dynamic coupling relationship has formed among "PV-Energy Storage-Grid," posing a severe challenge to system stability control.

[0003] The existing optimization techniques for the rotational inertia (including physical and virtual inertia) and damping coefficient parameters of grid-type photovoltaic-storage systems have several shortcomings that urgently need to be addressed: First, the characterization of multi-device coupling characteristics is inaccurate. Traditional methods often use first-order linear models, designing parameters independently for each single device, ignoring the nonlinear dynamic interaction between devices, which can easily lead to system coupling oscillations. Second, parameter selection lacks dynamic coordination. Inertia and damping coefficients are mostly fixed values, relying solely on experience or single-condition optimization, which cannot adapt to dynamic scenarios such as grid load fluctuations and changes in photovoltaic output. Third, the optimization algorithm has poor adaptability. Traditional optimization algorithms (fixed-weight particle swarm optimization) have inherent defects such as fixed inertia weights and static learning factors. They do not combine system coupling with dynamic adjustment of the search strategy, are prone to getting trapped in local optima, and have coarse constraint handling, resulting in insufficient optimization efficiency and engineering feasibility. Fourth, there is a lack of a sound adaptive adjustment mechanism. The optimized parameters are directly injected into the control loop, failing to avoid system shocks caused by parameter mutations, and failing to make dynamic corrections based on real-time operating status, thus failing to achieve closed-loop coordination of "state perception - parameter update - stable control".

[0004] The aforementioned problems make it difficult for existing technologies to meet the voltage-frequency stability requirements of the power grid in scenarios with a high proportion of renewable energy grid connection. There is an urgent need for a technical solution that can accurately characterize the coupling relationship of multiple devices, dynamically optimize parameters, and achieve adaptive adjustment in order to improve the grid connection stability and dynamic response robustness of grid-connected photovoltaic-storage systems. Summary of the Invention

[0005] To address the problems in existing parameter optimization for grid-type photovoltaic-storage systems, such as inaccurate characterization of multi-device coupling characteristics, lack of dynamic coordination in parameter selection, poor adaptability of optimization algorithms, imperfect adjustment mechanisms, and lack of verification systems, this invention provides a parameter optimization method for adaptive damping inertia adjustment in grid-type photovoltaic-storage systems. This method involves establishing a high-precision multi-device coupling model, constructing a multi-objective optimization function, designing an improved optimization algorithm incorporating coupled dynamic feedback, building an adaptive adjustment mechanism, and verifying the effect through simulation and testing. This achieves dynamic optimization and adaptive adjustment of rotational inertia and damping coefficient, ultimately suppressing multi-device coupling oscillations, improving the system's voltage-frequency coordinated stability and dynamic response robustness, and providing technical support for high-proportion grid-connected renewable energy systems.

[0006] The technical solution adopted in the parameter optimization method for adaptive damped inertia adjustment of a grid-type photovoltaic energy storage system of the present invention is as follows:

[0007] A parameter optimization method for adaptive damping inertia adjustment of a grid-type photovoltaic energy storage system includes a truss main body, a high-voltage cable laying unit, and a low-voltage cable laying unit; the truss main body is an overhead support structure, and the high-voltage cable laying unit and the low-voltage cable laying unit are arranged in layers along the height direction of the truss main body, and the height of the high-voltage cable laying unit is higher than the height of the low-voltage cable laying unit.

[0008] A further improvement of the technical solution of the present invention is that: the high-voltage cable laying unit is located in the upper area of ​​the truss body, the low-voltage cable laying unit is located in the lower area of ​​the truss body, and a preset heat dissipation distance is reserved between the high-voltage cable laying unit and the low-voltage cable laying unit to ensure the heat dissipation requirements of the high and low voltage cables during operation.

[0009] A further improvement of the technical solution of the present invention is that: the preset heat dissipation distance is not less than 0.5m, and both the high-voltage cable laying unit and the low-voltage cable laying unit are equipped with cable bearing components.

[0010] A further improvement of the technical solution of the present invention is that: the cable bearing components of the high voltage cable laying unit and the low voltage cable laying unit are provided with a first cable fixing component. The first cable fixing component includes a support bracket and a positioning clamp. The interval between two adjacent support brackets is 1m. The positioning clamp is a non-magnetic cable fixing clamp. Two adjacent non-magnetic cable fixing clamps are arranged on the support bracket at a interval of 2m.

[0011] A further improvement of the technical solution of the present invention is that: the cable carrying assembly further includes a second cable fixing assembly, the second cable fixing assembly includes a binding member, and the inner side of the binding member is provided with a buffer pad for avoiding hard friction between the cable and the binding member.

[0012] A further improvement of the technical solution of the present invention is that: the high-voltage cable laying unit further includes a high-voltage cable lead-out structure, which is connected to the outgoing side of the 220kV power distribution device. After the high-voltage cable is led out through the high-voltage cable lead-out structure, it extends along the upper area of ​​the truss body and is mounted on the corresponding cable bearing component.

[0013] A further improvement of the technical solution of the present invention is that: the low-voltage cable laying unit further includes a low-voltage cable lead-out structure, the low-voltage cable lead-out structure is connected to the outer cable trench of column A, the low-voltage cable enters the low-voltage cable lead-out structure through the outer cable trench of column A, and after being led out by the low-voltage cable lead-out structure, it extends along the lower area of ​​the truss body and is mounted on the corresponding cable bearing component.

[0014] A further improvement of the technical solution of the present invention is that it also includes a cable identification system, wherein the cable identification system includes a number of identification elements that record the unique number, specifications and usage information of the cable, and the number of identification elements are respectively set on each high-voltage and low-voltage cable.

[0015] The technological advancements achieved by this invention due to the adoption of the above technical solutions are as follows:

[0016] This invention breaks through the limitations of existing first-order linear models and constructs a second-order nonlinear coupling model covering cross-device interactions of photovoltaics and energy storage, accurately characterizing the bidirectional coupling mechanism of "DC voltage-power-frequency". It abandons the single-index optimization approach and integrates multiple objectives such as frequency deviation, AC / DC voltage fluctuation, equipment loss, and coupling oscillation suppression to achieve synergistic optimization of grid stability and equipment lifespan, solving the defects of traditional methods that are "inaccurate in characterization and neglect one aspect for another".

[0017] This invention addresses the inherent shortcomings of traditional PSO algorithms by innovatively designing coupling strength adaptive inertial weights and distance-constraint-aware learning factors, incorporating coupling dynamic feedback and constraint-aware repair strategies, thereby improving both search accuracy and the proportion of feasible solutions. A closed-loop adaptive adjustment mechanism of "state awareness-parameter decision-smooth update-feedback correction" is established to avoid parameter abrupt shocks, significantly improving the system's dynamic response speed and maintaining high robustness under complex disturbance conditions. Attached Figure Description

[0018] Figure 1 This is a system topology diagram of the parameter optimization method for adaptive damping inertia adjustment of a grid-type photovoltaic energy storage system according to the present invention.

[0019] Figure 2 The following is a flowchart of an improved particle swarm optimization algorithm for an adaptive damping inertia adjustment parameter optimization method for a grid-type photovoltaic energy storage system according to the present invention.

[0020] Figure 3This is a control block diagram of a parameter optimization method for adaptive damping inertia adjustment in a grid-type photovoltaic energy storage system according to the present invention.

[0021] Figure 4 Comparison of frequency recovery simulation results under grid-connected scenarios for grid-connected photovoltaic-storage systems;

[0022] Figure 5 Comparison of voltage recovery simulation results under grid-connected scenarios for grid-connected photovoltaic-storage systems;

[0023] Figure 6 A comparison chart of simulation results of active power output fluctuations in a grid-connected photovoltaic-storage system scenario. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings. In the following description, descriptions of formula structures and techniques are omitted to avoid unnecessarily obscuring the concepts of this invention.

[0025] Example 1

[0026] like Figure 1-3 As shown in the figure, this embodiment discloses a parameter optimization method for adaptive damping inertia adjustment of a grid-type photovoltaic energy storage system, including the following steps:

[0027] S1. Establish a dynamic model of multi-device coupling in a grid-type photovoltaic-storage system.

[0028] Photovoltaics is based on the principle of capacitor charging and discharging. The energy balance equation on the DC side is:

[0029]

[0030] In the formula, For DC side capacitors; It is a DC voltage; This is the photovoltaic output current; Inverter input current.

[0031] By adjusting the modulation ratio Injecting photovoltaic virtual inertia and photovoltaic damping After responding to changes in grid frequency, the inverter power equation containing virtual inertia is:

[0032]

[0033]

[0034] In the formula, This refers to the output power of the photovoltaic inverter. The voltage at point PCC; Power angle of photovoltaic inverter; This refers to the equivalent reactance of a photovoltaic inverter. The modulation ratio reference value; This is the dynamic adjustment amount for the photovoltaic modulation ratio; The virtual inertia modulation ratio coefficient of the photovoltaic system; This represents the photovoltaic damping modulation ratio coefficient.

[0035] Energy storage is based on the VSG principle, simulating the rotor motion of a synchronous generator. The core equation is:

[0036]

[0037] In the formula: For energy storage virtual inertia; This is the reference power for energy storage. This refers to the AC output power of the energy storage system. This is the energy storage damping coefficient; This is the rated angular velocity of the power grid.

[0038] The energy balance equation for the parallel bus at the DC end of the energy storage is:

[0039]

[0040] In the formula, Total bus capacitance; Common DC bus voltage; It is the energy storage DC current.

[0041]

[0042] In the formula, For energy storage DC-side power, and AC-side power Satisfy the transformation relationship; For energy storage converter efficiency.

[0043] By combining the above sub-models and linearizing them with a small perturbation, a system-level state-space model is constructed. The state vector X and control variables are defined. Perturbation vector :

[0044]

[0045]

[0046]

[0047] In the formula, Disturbance to photovoltaic output; This is a load power disturbance.

[0048] The state equations for coupling are:

[0049]

[0050] In the formula, This is the system matrix, which is related to the control variables and determines the dynamic characteristics of the system.

[0051] The output equation is:

[0052]

[0053] In the formula, This refers to the power grid frequency deviation. This represents the voltage deviation at the PCC point.

[0054] S2. Using the moment of inertia and damping coefficient as optimization variables, construct a multi-objective optimization function and constraints.

[0055] Construct a multi-objective optimization function. Define the moment of inertia vector J and the damping coefficient vector D:

[0056]

[0057]

[0058] Based on the S1 coupling model, a minimum objective function is constructed to comprehensively quantize the frequency deviation. Voltage deviation Equipment loss and multi-device coupled oscillation The expression is as follows:

[0059]

[0060] Weighting coefficient constraints:

[0061]

[0062] It can be dynamically adjusted according to the power grid operation scenario. All objective items have been normalized to ensure uniformity of dimensions and avoid a single objective dominating the optimization result.

[0063] To ensure the engineering feasibility of the optimized parameters, the following constraints are set based on the equipment's physical characteristics, national standards, and the S1 model:

[0064] Optimize the constraint variables; the range of values ​​for constraints J and D is:

[0065]

[0066] To match the voltage withstand range of the DC-DC converter, the photovoltaic DC voltage is constrained:

[0067]

[0068] Energy storage operation constraints:

[0069]

[0070] In the formula, This is the rated power of the energy storage.

[0071] To avoid voltage distortion caused by overmodulation, the inverter modulation ratio must be constrained.

[0072]

[0073] System operating constraints, frequency deviation constraints:

[0074]

[0075] Frequency change rate constraint (minimum requirement for power grid inertia support):

[0076]

[0077] Voltage deviation constraint:

[0078] .

[0079] S3. Design an improved particle swarm optimization (PSO) algorithm incorporating coupled dynamic feedback. The process of this improved PSO algorithm is as follows: Figure 2 As shown: After the algorithm starts, it first initializes the parameters, completes the setting of particle encoding, particle swarm size and number of iterations, and generates the initial position and velocity of the particles; then it calculates the initial fitness and records the individual optimal and global optimal positions; then it performs adaptive parameter adjustment based on the system coupling strength and particle search state, and updates the particle velocity and position. For particle positions that exceed the constraints, it performs constraint-aware repair through "boundary adsorption + slight perturbation"; then it recalculates the fitness and updates the optimal position, and then determines whether the convergence condition is met. If it is met, it outputs the optimal parameters and ends the process. If it is not met, it determines whether the iteration termination condition has been reached. If it has not been reached, it returns to the adaptive parameter adjustment stage to continue iterating. If it has been reached, it outputs the optimal parameters and ends the process.

[0080] Particle encoding and initialization. Each particle corresponds to a complete set of optimization variables (J, D). Real-number encoding is used to ensure search continuity. The particle position vector is defined as:

[0081]

[0082] In the formula, N is the particle swarm size, and the value range of each component strictly matches the optimization variable constraints.

[0083] Initialize particle positions to avoid initial particle aggregation: for each component ,according to Generate, and simultaneously pre-screen through equipment operation constraints to eliminate initially infeasible solutions; particle velocity Initialization range is , To ensure a stable initial speed.

[0084] Adaptive inertia weighting for coupling strength. Design of adaptive inertia weighting. :

[0085]

[0086] In the formula, , The range of basic weights; For the fitness of particles; The average fitness of the particle swarm; , These represent the maximum and minimum fitness of the population, respectively. It is a quantitative index of system coupling strength; the larger the value, the more intense the system coupling oscillation. This is the coupling strength weighting coefficient, which balances the influence of the population state and coupling characteristics.

[0087] The adaptive logic is: when there is strong coupling or high group diversity hour, Approaching Enhance global exploration capabilities and avoid getting trapped in local optima; when weakly coupled or group convergence hour, Approaching This accelerates local convergence and improves search accuracy.

[0088] Coupled dynamic feedback velocity update formula. The velocity update formula is:

[0089]

[0090] In the formula, These are called basic terms, which retain inertia characteristics and are determined by adaptive weights. Adjustment; This is called the individual cognitive term, which guides the particle toward its own optimal solution. search, An adaptive individual learning factor; Called search, For adaptive global learning factors; This is called the coupled dynamic feedback term, which transforms the coupled state of the system into a velocity adjustment quantity. For feedback coefficients, This is the coupling feedback vector.

[0091] Adaptive learning factor Design, incorporating particle constraint satisfaction ( (Indicates that the constraints are fully satisfied) and the distance from the particle to the global optimum. :

[0092]

[0093] Dynamic adjustment:

[0094]

[0095]

[0096] In the formula, When the particle is large (far from the optimal solution), Enlarge Reduce, enhance individual exploration; When the particle size is small (close to the optimal solution), Reduce Increase and strengthen overall utilization; When the particle is close to the constraint boundary, reduce and This avoids generating infeasible solutions.

[0097] Constraint-aware location update and feasible solution repair. Location update formula:

[0098]

[0099] For the updated Verify each constraint (optimization variable constraints, equipment operation constraints, system operation constraints). If any constraint is violated, perform the following operations:

[0100] To optimize variable constraints, the variable is "attracted" to the nearest constraint boundary, forcing the feasibility requirement to be met. Then, a small-scale, random perturbation is applied to the variable attracted to the boundary to prevent multiple particles from gathering at the same boundary point, which would cause the search space to shrink and get trapped in local optima.

[0101]

[0102] Equipment operation constraints: Corrected by adjusting the corresponding inertia J and damping D, the correction amount is positively correlated with the degree of constraint violation.

[0103] Fitness evaluation and convergence judgment. A multi-objective function F(J, D) is used as the fitness; the smaller the fitness, the better the particle. For the repaired feasible solution, an additional "constraint satisfaction reward term" is introduced.

[0104]

[0105] In the formula, The reward coefficient motivates particles to satisfy constraints.

[0106] The iteration stops when both of the following conditions are met: the change in the optimal fitness of the population. Feasible solution ratio (the proportion of feasible particles to the total number of particles) If the iteration count reaches the maximum value and convergence is still not achieved, the process is forcibly stopped and the current optimal feasible solution is output.

[0107] S4. Construct an adaptive adjustment mechanism.

[0108] Real-time system status awareness and quantitative assessment. The following core status variables are collected in real time through the PCC point monitoring unit, the common DC bus monitoring module, and local sensors on each device:

[0109] Grid-side conditions include real-time grid angular velocity. PCC point voltage Rate of change of frequency ;

[0110] Equipment-side conditions include common DC bus voltage Energy storage AC side output power Remaining energy storage capacity SOC(t);

[0111] Disturbances and coupling states include photovoltaic power output fluctuations. System coupling strength :

[0112]

[0113] To accurately determine the direction and magnitude of parameter adjustments, three types of quantitative evaluation indicators are defined, all of which have been normalized (value range [0,1]):

[0114] Disturbance intensity index :

[0115]

[0116] This formula comprehensively reflects the magnitude of the disturbance experienced by the system. This represents the upper limit of the rate of change of frequency. These are the rated power of the photovoltaic system, The larger the value, the stronger the disturbance.

[0117] Voltage stability index :

[0118]

[0119] The larger the value, the more severe the voltage fluctuation.

[0120] Coupled Oscillation Indicator The larger the value, the more obvious the coupled oscillation.

[0121] Dynamic decision-making for inertia and damping parameters. Based on the above quantitative evaluation indicators and combined with the multi-device collaborative selection principle of S3, a dynamic parameter decision-making model is constructed to correct the optimal parameters in real time. and To obtain the target parameter vector at the current time. and The specific decision-making formula is as follows:

[0122] Energy storage target inertia :

[0123]

[0124] As the dominant inertia of the system, its decision-making priority response to disturbance intensity and frequency dynamics. When At that time, an additional attenuation coefficient of 0.7 is introduced (to avoid over-discharge of stored energy). At that time, an additional gain factor of 1.2 is introduced (to fully utilize the energy storage capacity), and the following condition is always met. .

[0125] Photovoltaic target inertia :

[0126]

[0127] Adapts to fluctuations in photovoltaic power output and dynamic changes in DC voltage. When and hour, (Simplified control) to meet .

[0128] Energy storage target damping :

[0129]

[0130] The dominant system damping, response coupled oscillation and frequency deviation, when the system oscillation frequency hour, Pick (Enhanced low-frequency oscillation suppression), satisfying .

[0131] Photovoltaic target damping :

[0132]

[0133] Stable DC voltage, associated energy storage damping, and satisfying ,make sure .

[0134] A smooth parameter transition update mechanism. This is to avoid affecting the target parameter. and Sudden changes trigger system oscillations. A smooth transition function based on a first-order low-pass filter is designed to adjust the current operating parameters. and Gradually update to the target parameters.

[0135] For each inertia component The smooth update formula is:

[0136]

[0137] In the formula, The sampling period; This is the inertia smoothing factor (with a value range of 0.05 to 0.2).

[0138] Its dynamic adjustment logic is: strong disturbance hour, (Speed ​​up update) Weak perturbation hour, Slow updates, maintaining stability); under moderate disturbances, (Balancing speed and stability).

[0139] For each damping component The smooth update formula is:

[0140]

[0141] In the formula, This is the damping smoothing factor (with a value range of 0.1 to 0.3).

[0142] Its dynamic adjustment logic is: strong oscillation hour, (Accelerate and strengthen damping); weak oscillation hour, (Slow adjustment); During moderate oscillations, .

[0143] Control loop parameter injection and adaptive execution. Smoothly updated inertia parameters. , With damping parameters , The control loops for photovoltaic and energy storage are injected separately, and the control equations are updated in conjunction with the S1 device sub-model to achieve adaptive damping inertia adjustment:

[0144] The formula for the dynamic adjustment of the photovoltaic modulation ratio leads to the adaptive modulation ratio control equation:

[0145]

[0146] In the formula, This is the damping-modulation ratio coefficient.

[0147] Energy storage adaptive VSG control equations:

[0148]

[0149] Closed-loop feedback correction. The adaptive adjustment mechanism achieves continuous optimization through closed-loop feedback. Specifically, after each parameter injection, the system output vector is monitored in real time. and Calculate the feedback correction coefficient When the system deviation is large At that time, the smoothing factor was lowered. and The value of (to speed up parameter updates) and the adjustment range of the target parameter are increased; when the system deviation is small At this time, the smoothing factor is increased (to slow down updates) to maintain parameter stability.

[0150] S5. Verify the effect of the method in this embodiment on the inertia injection and voltage-frequency coordinated stability control in grid-connected control through simulation or testing, and monitor the bidirectional coupling ability to suppress voltage fluctuations and frequency oscillations to improve the dynamic response and stability of the system.

[0151] The simulation model of the grid-connected photovoltaic-storage system in this embodiment is built on the Matlab / Simulink simulation platform, using a "photovoltaic-storage centralized coupling" topology (such as...). Figure 1 The diagram shows the principle framework, which reproduces the equipment characteristics and coupling mechanism in a real grid-connected scenario. The total simulation time is set to 20 seconds, and the experimental process design covers multiple typical disturbances: a large initial load is introduced at t=5 seconds, photovoltaic output fluctuations are applied at t=10 seconds, and the remaining load is introduced at t=15 seconds, which fully simulates the complex working conditions of the entire grid connection process. The system stability and control effect are judged by monitoring the core electrical quantities.

[0152] like Figure 4As shown, after optimizing the inertia vector J and damping coefficient vector D using the method of this embodiment, the system's frequency dynamic characteristics under multiple disturbance superposition scenarios are significantly better than those of traditional methods. Before optimization, the superposition of multiple disturbances caused a significant frequency drop, and the recovery to the rated value took a long time with obvious oscillations. After optimization, thanks to the precise configuration of "energy storage-dominated inertia + photovoltaic cooperative inertia" and the parameter tuning of the improved PSO algorithm, the frequency drop amplitude was effectively controlled within a reasonable range, the recovery speed was greatly improved, and the fluctuation curve was smooth and oscillating-free. This result shows that this embodiment achieves the globally optimal configuration of inertia and damping parameters through the optimization algorithm coupled with dynamic feedback, effectively improving the system's active support capability for frequency deviation, avoiding the frequency recovery lag problem caused by the imbalance of inertia coordination among multiple devices, and demonstrating the frequency stability control advantages of the method under complex disturbances in traditional control.

[0153] like Figure 5 As shown, the optimized system exhibits significantly improved stability of the common DC bus voltage, successfully achieving effective decoupling between DC voltage and grid frequency. At t=5s, when a large-capacity load is applied, before the photovoltaic output responds, the energy storage rapidly releases power through adaptive damping, reducing the maximum voltage drop on the DC bus and quickly restoring it to steady state. At t=10s, when photovoltaic output fluctuates, the photovoltaic virtual damping dynamically adjusts with the voltage deviation, suppressing the voltage oscillation amplitude. During subsequent load disturbances, the bus voltage remains stable without overshoot. Compared to traditional methods, the optimized DC voltage fluctuation amplitude is significantly reduced, fully validating the effectiveness of the voltage deviation term and coupled oscillation suppression term in the multi-objective optimization function of this embodiment, providing a solid guarantee for stable power transmission.

[0154] like Figure 6 As shown, the optimized system exhibits significantly enhanced synergy in active power output between photovoltaic (PV) and energy storage, effectively suppressing multi-device coupling oscillations. During initial load switching, energy storage rapidly compensates for power gaps, preventing sudden power surges. In subsequent load switching processes, the energy storage damping coefficient adjusts synchronously, offsetting the impact of various disturbances on the grid. PV power is output stably in maximum power point tracking mode, without additional oscillations due to DC voltage fluctuations or frequency disturbances. Compared to traditional methods, the optimized system shows a significantly reduced overall power fluctuation amplitude and no significant delay in multi-device power response. This verifies the scientific validity of the "energy storage-led, PV-coordinated" parameter selection principle in this embodiment, as well as the effectiveness of the adaptive adjustment mechanism in suppressing power coupling oscillations, ensuring the stability of grid-connected power and power quality.

[0155] Example 2

[0156] This embodiment, based on Embodiment 1, adjusts the "fixed-weight multi-objective optimization function" to a dynamic weight function, with the weights adjusted according to the real-time perturbation intensity. The adjustment is as follows:

[0157]

[0158] Applicable scenarios: Complex operating conditions with multiple disturbances superimposed; dynamic weights can adaptively switch performance priorities under different disturbance intensities.

[0159] Example 3

[0160] This embodiment replaces the "improved PSO algorithm incorporating dynamic feedback of optical-storage coupling" with an improved genetic algorithm incorporating optical-storage coupling feedback, based on Embodiment 1. The specific implementation is as follows:

[0161] Encoding method: Real number encoding is used, with each chromosome corresponding to a set of photovoltaic virtual inertia. Energy storage virtual inertia Photovoltaic virtual damping Energy storage virtual damping ;

[0162] Genetic operation: The crossover operator uses arithmetic crossover, and the crossover probability is based on the optical-storage coupling strength. Dynamic adjustment; the mutation operator employs Gaussian mutation, and the variable-asynchronous long-term correlated coupling oscillation index... ;

[0163] Fitness and constraint handling: The multi-objective optimization function and the "boundary adsorption + slight perturbation" strategy are the same as those in the main implementation.

[0164] Applicable scenarios: Large-scale distributed optical energy storage grid-connected systems (with a large number of devices and more complex coupling relationships). The improved global search capability of the genetic algorithm can adapt to the needs of parameter optimization on a larger scale.

[0165] The core objective of this invention is to systematically improve upon the technical shortcomings of existing grid-type photovoltaic-storage systems, such as inaccurate characterization of multi-device coupling, poor adaptability of parameter optimization algorithms, and rigid adjustment mechanisms. Existing technologies generally neglect the cross-device nonlinear coupling effects between photovoltaic, energy storage, and the grid, leading to a chain reaction of frequency oscillations, voltage fluctuations, and power imbalances under disturbances. Traditional optimization algorithms lack dynamic coupling awareness, making it difficult to obtain globally optimal parameters; and parameter adjustments are mostly fixed-value injections, unable to adapt to dynamic operating conditions. To address these problems, this invention proposes a complete technical solution: "high-precision coupling modeling - multi-objective optimization - improved algorithm solution - adaptive adjustment." This involves establishing a second-order nonlinear coupling model to accurately capture cross-device dynamic interactions; constructing a multi-objective function encompassing system stability, device losses, and coupling suppression to achieve multi-dimensional performance synergy; employing an improved PSO algorithm incorporating coupled dynamic feedback to improve parameter search accuracy and efficiency; and building a closed-loop adaptive adjustment mechanism to achieve smooth parameter updates and dynamic corrections.

[0166] Ultimately, simulations verified that the proposed solution not only suppressed the transmission of multi-device coupling disturbances at the source and improved the dynamic response speed and robustness of the grid-connected system, but also took into account the operating losses and service life of photovoltaic and energy storage. This provides reliable technical support and engineering reference for the safe and stable operation of distribution networks and microgrids with a high proportion of grid-connected new energy sources.

[0167] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the concept and scope of the present invention. Various modifications and improvements made to the technical solutions of the present invention by those skilled in the art without departing from the inventive concept should fall within the protection scope of the present invention. All technical contents for which protection is sought in this invention are fully described in the claims.

Claims

1. A parameter optimization method for adaptive damped inertia adjustment of a grid-type photovoltaic energy storage system, characterized in that, Includes the following steps: S1. Establish a dynamic model of multi-device coupling in a grid-type photovoltaic-storage system to accurately characterize the cross-device nonlinear coupling mechanism between photovoltaic, energy storage and grid, laying a theoretical foundation for parameter optimization; S2. Construct a multi-objective optimization function and constraints, using moment of inertia and damping coefficient as optimization variables, taking into account system stability, equipment loss and coupling suppression, and clarifying the feasible region of parameters; S3. Design an improved particle swarm optimization (PSO) algorithm incorporating coupled dynamic feedback to solve for the optimal parameter combination; S4. Establish an adaptive adjustment mechanism, dynamically correct and smoothly update parameters based on the real-time status of the system, and inject them into the control loop of each device to achieve adaptive damping inertia adjustment. S5. Through simulation analysis and actual testing, the control effect, robustness and engineering applicability of the method are fully verified.

2. The parameter optimization method for adaptive damped inertia adjustment of a grid-type photovoltaic energy storage system according to claim 1, characterized in that, In step S1, establishing a dynamic model of multi-device coupling in a network-type optical-storage system includes the following steps: S1.

1. Photovoltaics are based on the principle of capacitor charging and discharging. The DC-side energy balance equation is: ； In the formula, For DC side capacitors; It is a DC voltage; This is the photovoltaic output current; Inverter input current; By adjusting the modulation ratio Injecting photovoltaic virtual inertia and photovoltaic damping After responding to changes in grid frequency, the inverter power equation containing virtual inertia is: ； ； In the formula, This refers to the output power of the photovoltaic inverter. The voltage at point PCC; Power angle of photovoltaic inverter; This refers to the equivalent reactance of a photovoltaic inverter. The modulation ratio reference value; This is the dynamic adjustment amount for the photovoltaic modulation ratio; The virtual inertia modulation ratio coefficient of the photovoltaic system; This is the photovoltaic damping modulation ratio coefficient; S1.2 Energy storage is based on the VSG principle, simulating the rotor motion of a synchronous generator. The core equation is: ； In the formula: For energy storage virtual inertia; This is the reference power for energy storage. This refers to the AC output power of the energy storage system. This is the energy storage damping coefficient; The rated angular velocity of the power grid; The energy balance equation for the parallel bus at the DC end of the energy storage is: ； In the formula, Total bus capacitance; Common DC bus voltage; It is a DC current for energy storage; ； In the formula, For energy storage DC-side power, and AC-side power Satisfy the transformation relationship; For energy storage converter efficiency; S1.

3. Combining the above sub-models and linearizing them with small perturbations, construct a system-level state-space model; define the state vector X and control variables. Perturbation vector : ； ； ； In the formula, Disturbance to photovoltaic output; For load power disturbance; The state equations for coupling are: ； In the formula, This is the system matrix, which is related to the control variables and determines the dynamic characteristics of the system. The output equation is: ； In the formula, This refers to the power grid frequency deviation. This represents the voltage deviation at the PCC point.

3. The parameter optimization method for adaptive damped inertia adjustment of a grid-type photovoltaic energy storage system according to claim 1, characterized in that, In step S2, the steps of constructing the multi-objective optimization function and constraints, using the moment of inertia and damping coefficient as optimization variables, include: S2.1 Construct a multi-objective optimization function; define the moment of inertia vector J and the damping coefficient vector D: ； ； Based on the S1 coupling model, a minimum objective function is constructed to comprehensively quantize the frequency deviation. Voltage deviation Equipment loss and multi-device coupled oscillation The expression is as follows: ； Weighting coefficient constraints: ； It can be dynamically adjusted according to the power grid operation scenario; all target items have been normalized to ensure uniformity of dimensions and avoid a single target dominating the optimization result. S2.2 To ensure the engineering feasibility of the optimized parameters, based on the equipment's physical characteristics, national standard requirements, and model S1, the following constraints are set: Optimize the constraint variables; the range of values ​​for constraints J and D is: ； To match the voltage withstand range of the DC-DC converter, the photovoltaic DC voltage is constrained: ； Energy storage operation constraints: ； In the formula, Rated power of energy storage; To avoid voltage distortion caused by overmodulation, the inverter modulation ratio must be constrained. ； System operating constraints, frequency deviation constraints: ； Frequency change rate constraint: ； Voltage deviation constraint: 。 4. The parameter optimization method for adaptive damped inertia adjustment of a grid-type photovoltaic energy storage system according to claim 1, characterized in that, In S3, the operation steps of the improved particle swarm optimization (PSO) algorithm incorporating coupled dynamic feedback are as follows: S3.1 Particle Encoding and Initialization; Each particle corresponds to a complete set of optimization variables (J, D). Real number encoding is used to ensure search continuity. The particle position vector is defined as: ； In the formula, N is the particle swarm size, and the value range of each component strictly matches the optimization variable constraints; Initialize particle positions to avoid initial particle aggregation: for each component ,according to Generate, and simultaneously pre-screen through equipment operation constraints to eliminate initially infeasible solutions; particle velocity Initialization range is Ensure a stable initial velocity; S3.2, Adaptive inertia weighting for coupling strength; design of adaptive inertia weighting. : ； In the formula, , The range of basic weights; For the fitness of particles; The average fitness of the particle swarm; , These represent the maximum and minimum fitness of the population, respectively. It is a quantitative index of system coupling strength; the larger the value, the more intense the system coupling oscillation. The coupling strength weighting coefficient balances the influence of population state and coupling characteristics; The adaptive logic is: when or hour, Approaching Enhance global exploration capabilities and avoid getting trapped in local optima; when or hour, Approaching This accelerates local convergence and improves search accuracy. S3.3, Coupled Dynamic Feedback Velocity Update Formula; The velocity update formula is: ； In the formula, These are called basic terms, which retain inertia characteristics and are determined by adaptive weights. Adjustment; This is called the individual cognitive term, which guides the particle toward its own optimal solution. search, An adaptive individual learning factor; Called search, For adaptive global learning factors; This is called the coupled dynamic feedback term, which transforms the coupled state of the system into a velocity adjustment quantity. For feedback coefficients, This is the coupling feedback vector; Adaptive learning factor Design, incorporating particle constraint satisfaction Distance from the particle to the global optimum : ； Dynamic adjustment: ； ； In the formula, When it is big, Enlarge Reduce, enhance individual exploration; Hour, Reduce Increase and strengthen overall utilization; When low, reduce and To avoid generating infeasible solutions; S3.4, Constraint-Aware Position Update and Feasible Solution Repair; Position Update Formula: ； For the updated Verify each constraint one by one. If any constraint is violated, perform the following operations: Optimize variable constraints by "attracting" the variable to the nearest constraint boundary, forcing it to meet feasibility requirements. Then, apply a small-range, random perturbation to the variable attracted to the boundary to prevent multiple particles from gathering at the same boundary point, which would cause the search space to shrink and get trapped in local optima. ； Equipment operation constraints: Corrected by adjusting the corresponding inertia J and damping D, the correction amount is positively correlated with the degree of constraint violation; S3.5 Fitness Evaluation and Convergence Judgment; A multi-objective function F(J, D) is used as the fitness; the smaller the fitness, the better the particle; For the repaired feasible solution, an additional "constraint satisfaction reward term" is introduced: ； In the formula, The reward coefficient motivates particles to satisfy constraints; The iteration stops when both of the following conditions are met: the change in the optimal fitness of the population. ; proportion of feasible solutions If the iteration count reaches the maximum value and convergence is still not achieved, the process is forcibly stopped and the current optimal feasible solution is output.

5. The parameter optimization method for adaptive damped inertia adjustment of a grid-type photovoltaic energy storage system according to claim 1, characterized in that: In step S4, the step of constructing the adaptive adjustment mechanism includes: S4.1 Real-time System Status Awareness and Quantitative Assessment; The following core status variables are collected in real time through the PCC point monitoring unit, the common DC bus monitoring module, and local sensors of each device: Grid-side conditions include real-time grid angular velocity. PCC point voltage Rate of change of frequency ; Equipment-side conditions include common DC bus voltage Energy storage AC side output power Remaining energy storage capacity ; Disturbances and coupling states include photovoltaic power output fluctuations. System coupling strength : ； To accurately determine the direction and magnitude of parameter adjustments, three types of quantitative evaluation indicators are defined, and all indicators have been normalized: Disturbance intensity index : ； This formula comprehensively reflects the magnitude of the disturbance experienced by the system. This represents the upper limit of the rate of change of frequency. These are the rated power of the photovoltaic system, The larger the value, the stronger the disturbance; Voltage stability index : ； The larger the value, the more severe the voltage fluctuation; Coupled Oscillation Indicator The larger the value, the more pronounced the coupled oscillation; S4.2 Dynamic Decision-Making of Inertia and Damping Parameters: Based on the above quantitative evaluation indicators and combined with the multi-device collaborative selection principle of S3, a dynamic decision-making model for parameters is constructed to correct the optimal parameters in real time. and To obtain the target parameter vector at the current time. and The specific decision-making formula is as follows: Energy storage target inertia : ； As the dominant inertia of the system, its decision-making priority response to disturbance intensity and frequency dynamics; when When, an additional attenuation coefficient of 0.7 is introduced, when At that time, an additional gain coefficient of 1.2 is introduced, and the condition is always satisfied. ； Photovoltaic target inertia : ； Adapts to fluctuations in photovoltaic output and dynamic changes in DC voltage; when and hour, ,satisfy ； Energy storage target damping : ； The dominant system damping, response coupled oscillation and frequency deviation, when the system oscillation frequency hour, Pick ,satisfy ; Photovoltaic target damping : ； Stable DC voltage, associated energy storage damping, and satisfying ,make sure ; S4.3, Parameter smooth transition update mechanism; to avoid target parameter and Sudden changes trigger system oscillations. A smooth transition function based on a first-order low-pass filter is designed to adjust the current operating parameters. and Gradually update to the target parameters; For each inertia component The smooth update formula is: ； In the formula, The sampling period; This is the inertia smoothing factor; Its dynamic adjustment logic is as follows: hour, ; hour, During moderate disturbances, ; For each damping component The smooth update formula is: ； In the formula, This is the damping smoothing factor, with a value range of 0.1 to 0.3; Its dynamic adjustment logic is as follows: hour, ; hour, During moderate oscillations, ; S4.4, Control loop parameter injection and adaptive execution; smooth update of inertia parameters , With damping parameters , The control loops for photovoltaic and energy storage are injected separately, and the control equations are updated in conjunction with the S1 device sub-model to achieve adaptive damping inertia adjustment: The formula for the dynamic adjustment of the photovoltaic modulation ratio leads to the adaptive modulation ratio control equation: ； In the formula, This is the damping-modulation ratio coefficient; Energy storage adaptive VSG control equations: ； S4.5 Closed-loop feedback correction; The adaptive adjustment mechanism achieves continuous optimization through closed-loop feedback. The specific logic is as follows: After each parameter injection, the system output vector is monitored in real time. and Calculate the feedback correction coefficient ,when At that time, the smoothing factor was lowered. and The value of is adjusted by increasing the adjustment range of the target parameter; when the system deviation is small. At that time, the smoothing factor is increased to maintain parameter stability.

Images