Multi-time-scale unified modeling and equivalence method and system for network construction type energy storage system

Abstract

The invention provides a multi-time-scale unified modeling and equivalence method and system for a network construction type energy storage system, and belongs to the technical field of power system modeling. The method comprises the following steps: classifying full-state variables of the network construction type energy storage system according to a dominant dynamic time scale, and constructing a three-layer nested differential manifold based on the time scale classification to obtain a geometric state space; establishing a generalized Lagrangian unified kinetic equation of the network construction type energy storage system, carrying out collaborative optimization on the parameterized equivalent model, and deploying a hierarchical parallel adaptive updating mechanism to enable the parameterized equivalent model to continuously evolve in a full life cycle, and realizing multi-time-scale unified modeling and physical consistency equivalence of the network-forming type energy storage system. According to the method, a unified kinetic equation is established based on a generalized Lagrange framework, analytic calculation and quantitative mapping are performed on cross-scale coupling by utilizing Lie bracket operation, and accurate quantitative description of a dynamic coupling mechanism between different time scales is realized.

Description

Technical Field

[0001] This invention relates to the field of power system modeling technology, specifically to a unified modeling and equivalence method and system for multi-timescale energy storage systems. Background Technology

[0002] With the increasing integration of high-proportion renewable energy into the grid and the deepening of power electronics, the dynamic characteristics of power systems are becoming increasingly complex, placing higher demands on the accuracy of key equipment models for stability analysis and operation control. Grid-based energy storage systems, as core equipment supporting next-generation grid stability, can autonomously construct grid voltage and frequency, providing rapid active and reactive power support and virtual inertia, thus their importance is increasingly prominent. Correspondingly, modeling and simulation technologies for grid-based energy storage systems have evolved from simple to complex. Early research focused on the electromagnetic transient model of its converter and the quasi-steady-state model based on active power frequency and reactive power voltage droop. To balance simulation efficiency and accuracy, multi-rate simulation and interface technologies have been applied, specifically calculating electromagnetic and electromechanical transient processes separately at different time lengths in the simulation. In recent years, detailed control models encompassing rotor motion equations and voltage regulation have become a research hotspot to more realistically characterize the electromechanical transients of "virtual synchronous machines." Simultaneously, with increasing attention to the economic efficiency and reliability of energy storage systems throughout their entire lifecycle, electrothermal coupling models combining slow dynamics (seconds and above) with electrical dynamics are gradually entering the exploratory stage. However, the lack of a unified theoretical framework in these approaches leads to the inability to accurately describe and quantify the coupling mechanisms between dynamics at different time scales, resulting in a loss of model predictive accuracy when analyzing complex system-level problems. Summary of the Invention

[0003] The purpose of this invention is to provide a unified modeling and equivalent method and system for multi-timescale models of grid-type energy storage systems, which can automatically maintain physical consistency and continuously evolve throughout the entire life cycle for power system simulation and analysis.

[0004] To achieve the above objectives, this invention provides a unified multi-timescale modeling and equivalence method for grid-type energy storage systems, comprising: classifying the full-state variables of the grid-type energy storage system according to the timescale of the dominant dynamics, and constructing a three-layer nested differential manifold based on the timescale classification to obtain a geometric state space; establishing the generalized Lagrange unified dynamic equation of the grid-type energy storage system based on the geometric state space, explicitly separating the dynamics of multiple timescales by introducing scale parameters, and quantifying the coupling relationship between dynamics of different timescales using Lie bracket operations; constructing a parameterized equivalence model according to the unified dynamic equation and the quantized coupling relationship, and co-optimizing the parameterized equivalence model; verifying the co-optimized parameterized equivalence model under full operating conditions, and deploying a hierarchical parallel adaptive update mechanism to enable the parameterized equivalence model to continuously evolve throughout its entire lifecycle, thereby achieving unified multi-timescale modeling and physical consistency equivalence of the grid-type energy storage system.

[0005] Optionally, the total state variables are classified into fast-scale state sets, medium-scale state sets, and slow-scale state sets according to the time scale of the dominant dynamic.

[0006] Optionally, the step of constructing a three-layer nested differential manifold based on the time scale classification to obtain the geometric state space includes: constructing corresponding submanifolds for the fast-scale state set, the mesoscale state set, and the slow-scale state set respectively; combining the submanifolds into a smooth product manifold to form a three-layer nested differential manifold; defining the ratio of the time constants of the fast-scale dynamics to the mesoscale dynamics as the first scale separation parameter, and defining the ratio of the time constants of the mesoscale dynamics to the slow-scale dynamics as the second scale separation parameter.

[0007] Optionally, the step of constructing a three-layer nested differential manifold based on the time scale classification to obtain a geometric state space further includes: defining a geometric structure on the geometric state space; the geometric structure includes a kinetic energy metric tensor, a canonical symplectic structure, and a Rayleigh dissipation function, wherein the kinetic energy metric tensor is configured to quantitatively characterize the inertial coupling strength between state variables through its matrix elements; the canonical symplectic structure is configured to uniquely determine multiple canonical coordinate pairs, each canonical coordinate pair corresponding to a pair of physical conservation quantities; the Rayleigh dissipation function is configured to characterize the energy loss in the grid-type energy storage system, the energy loss including switching conduction loss, virtual synchronizing machine damping loss, thermal conduction loss, and battery internal resistance loss.

[0008] Optionally, the explicit separation of multi-timescale dynamics by introducing scale parameters includes: using the first scale separation parameter and the second scale separation parameter as timescale transformation factors for corresponding dynamic terms, and introducing them into the generalized Lagrange unified dynamic equation; multiplying the first scale separation parameter by all generalized inertial terms in the equation related to mesoscale dynamics, and multiplying the second scale separation parameter by all generalized inertial terms in the equation related to slow-scale dynamics, to obtain the scale-reconstructed generalized Lagrange unified dynamic equation; and reconstructing the scale-reconstructed generalized Lagrange unified dynamic equation into three mutually... Coupled subsystems are used to achieve explicit separation of dynamics across multiple time scales. The three mutually coupled subsystems are a first time scale subsystem, a second time scale subsystem, and a third time scale subsystem. The first time scale subsystem represents the electromagnetic transient process dominated by the natural time constant of the grid-type energy storage system. The second time scale subsystem represents the electromechanical transient process and control dynamic process dominated by the time constant slowed down by the first time scale separation parameter. The third time scale subsystem represents the energy dispatch process and aging evolution process dominated by the time constant slowed down by the second time scale separation parameter.

[0009] Optionally, the quantification of coupling relationships between dynamics at different time scales using Lie bracket operations includes: defining fast-scale vector fields, mesoscale vector fields, and slow-scale vector fields for fast-scale state sets, mesoscale state sets, and slow-scale state sets, respectively, based on the generalized Lagrange unified dynamic equations; calculating a first Lie bracket between the fast-scale vector field and the mesoscale vector field, a second Lie bracket between the mesoscale vector field and the slow-scale vector field, and a third Lie bracket between the fast-scale vector field and the slow-scale vector field; quantifying the nonlinear coupling strength between fast-scale dynamics and mesoscale dynamics, between mesoscale dynamics and slow-scale dynamics, and between fast-scale dynamics and slow-scale dynamics, respectively, based on the magnitudes of the first, second, and third Lie brackets; and mapping each quantized nonlinear coupling strength to engineering parameters including coupling gain and coupling delay.

[0010] Optionally, the construction of the parameterized equivalent model and the collaborative optimization of the parameterized equivalent model include: constructing the parameterized equivalent model, which includes a first equivalent module, a second equivalent module, a third equivalent module, and a cross-scale coupling module; the first equivalent module corresponds to the fast-scale dynamics and includes a controlled voltage source and a frequency-varying impedance network; the second equivalent module corresponds to the mesoscale dynamics and is an extended virtual synchronous machine model; the third equivalent module corresponds to the slow-scale dynamics and includes time-varying parameters and a time-varying parameter battery thermal coupling model integrating aging prediction.

[0011] Optionally, the construction of the parameterized equivalent model and the collaborative optimization of the parameterized equivalent model further includes: performing step-by-step preliminary identification of the parameters of the parameterized equivalent model to obtain the initial parameter values ​​of each module of the parameterized equivalent model; using the coupling gain and coupling delay of the cross-scale coupling module as coupling constraints, and starting from the initial parameter values, performing collaborative optimization of the parameters of all modules; the collaborative optimization is to minimize the output error of the parameterized equivalent model under the condition of satisfying the coupling constraints.

[0012] Optionally, the constrained feedforward update layer in the hierarchical parallel adaptive update mechanism uses a constrained recursive estimation algorithm to update the parameters of the parameterized equivalent model online during the grid-connected operation of the grid-type energy storage system.

[0013] On the other hand, the present invention provides a unified modeling and equivalence system for multi-timescale grid-type energy storage systems, which is used to realize the unified modeling and equivalence method for multi-timescale grid-type energy storage systems. The system includes a control module, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program to realize the unified modeling and equivalence method for multi-timescale grid-type energy storage systems.

[0014] The aforementioned technical solution provides a unified geometric state space for multi-timescale dynamics by constructing a three-layer nested differential manifold, and establishes unified dynamic equations based on the generalized Lagrange framework, fundamentally establishing a unified theoretical framework for describing multi-scale dynamics. Within this framework, Lie bracket operations are used to perform analytical calculations and quantitative mappings of cross-scale couplings, thereby achieving a precise quantitative description of the coupling mechanisms between dynamics at different timescales. This overcomes the black-box nature of coupling mechanisms caused by traditional decoupling methods, enabling the maintenance of prediction accuracy and physical realism when analyzing system-level problems involving complex cross-scale interactions.

[0015] Other features and advantages of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0016] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the following detailed description to explain the invention, but do not constitute a limitation thereof. In the drawings:

[0017] Figure 1 This is a flowchart of a unified modeling and equivalence method for multi-timescale energy storage systems.

[0018] Figure 2 It is a path diagram for scale separation and coupling quantization of multi-scale dynamic equations of grid-type energy storage systems. Detailed Implementation

[0019] The following is in conjunction with the appendix Figure 1 -Appendix Figure 2 The specific implementation methods of the embodiments of the present invention will be described in detail below. It should be understood that the specific implementation methods described herein are only for illustrating and explaining the embodiments of the present invention, and are not intended to limit the embodiments of the present invention.

[0020] It should be noted that the acquisition, transmission, storage, use, and processing of data in the technical solution of this application all comply with the relevant provisions of national laws and regulations. In the embodiments of this application, certain existing industry solutions such as software, components, and models may be mentioned. These should be considered exemplary, intended only to illustrate the feasibility of implementing the technical solution of this application, and do not imply that the applicant has already used or necessarily used such solutions.

[0021] In the process of realizing this invention, the inventors of this application discovered that the prior art lacks a unified theoretical framework, which makes it impossible to accurately describe and quantify the coupling mechanism between dynamics at different time scales, thus resulting in a loss of model prediction accuracy when analyzing complex system-level problems.

[0022] Example 1

[0023] Reference Figures 1-2 This is the first embodiment of the present invention, which provides a unified multi-timescale modeling and equivalence method for a grid-type energy storage system, including:

[0024] S100: The full state variables of the grid-type energy storage system are classified according to the time scale of the dominant dynamic, and a three-layer nested differential manifold is constructed based on the time scale classification to obtain the geometric state space.

[0025] In the embodiments of this application, the total state variables are classified into fast-scale state sets, medium-scale state sets, and slow-scale state sets according to the time scale of the dominant dynamic.

[0026] In a preferred embodiment of this application, all independent state variables of the grid-type energy storage system are first collected to ensure that no key dynamic characterization parameters are omitted. Then, the state variables are classified into three levels according to their dominant time constants: fast-scale state sets, mesoscale state sets, and slow-scale state sets.

[0027] Fast-scale state set The time range is , It includes at least the switching duty cycle function of the grid-type energy storage system (the periodic average of the original switching function), the filter inductor current, capacitor voltage, and high-frequency components of the DC bus voltage of the grid-type energy storage system.

[0028] Mesoscale state set The time range is , It includes at least the phase-locked loop output angle and frequency of the grid-type energy storage system, the rotor angle and angular velocity of the virtual synchronous machine, and the integral state of the power control loop;

[0029] Slow-scale state set The time range is , It includes at least the battery's state of charge, average temperature, capacity decay status, and internal resistance growth status, and may also include the scheduling instructions of the grid-type energy storage system.

[0030] The time range of each state set is defined and divided based on the inherent time constant of the physical process it governs (power conversion hardware limits, control and synchronization dynamics, energy management and aging processes) and in conjunction with the engineering consensus on multi-scale analysis of power systems.

[0031] Preferably, to ensure the continuity of the state variables of the grid-type energy storage system, all discrete states (such as the switching function of the grid-type energy storage system) are processed into continuous states using the periodic averaging method or the Sigmoid function, thereby ensuring that the state space of the grid-type energy storage system satisfies the requirements of the grid-type energy storage system. Smoothness.

[0032] In the embodiments of this application, a three-layer nested differential manifold is constructed based on time scale classification to obtain a geometric state space, including: constructing corresponding submanifolds for the fast-scale state set, the mesoscale state set, and the slow-scale state set respectively; combining the submanifolds into a smooth product manifold to form a three-layer nested differential manifold; defining the ratio of the time constants of the fast-scale dynamics to the mesoscale dynamics as a first-scale separation parameter, and defining the ratio of the time constants of the mesoscale dynamics to the slow-scale dynamics as a second-scale separation parameter.

[0033] In a preferred embodiment of this application, a smooth product differential manifold M of the total state space of a grid-type energy storage system is constructed based on three types of state sets.

[0034]

[0035] in, , , These are the fast, medium, and slow-scale submanifolds of a grid-type energy storage system, all of which satisfy... Smoothness By processing the fast-scale state set after continuumization and smoothing... It is obtained by equipping a smooth differential structure. , Similarly, , , This forms a three-layer nested structure.

[0036] Furthermore, a hierarchical scale separation parameter is defined to quantify the time ratio relationship at different scales.

[0037] First-scale separation parameters ( =1ms is a typical fast-scale time constant for grid-type energy storage systems. =100ms is the typical time constant for medium-scale operations in a grid-type energy storage system, corresponding to the fast-to-medium-scale ratio of the grid-type energy storage system;

[0038] Second-scale separation parameters ( =1000ms is the typical time constant for slow scales), corresponding to the medium-slow scale ratio of grid-type energy storage systems;

[0039] Fast-slow composite scale ratio This corresponds to the comprehensive ratio between the fast and slow scales of a grid-type energy storage system.

[0040] In the embodiments of this application, a three-layer nested differential manifold is constructed based on time-scale classification to obtain a geometric state space. The method further includes: defining a geometric structure on the geometric state space; the geometric structure includes a kinetic energy metric tensor, a canonical symplectic structure, and a Rayleigh dissipation function, wherein the kinetic energy metric tensor is configured to quantitatively characterize the inertial coupling strength between state variables through its matrix elements; the canonical symplectic structure is configured to uniquely determine multiple canonical coordinate pairs, each canonical coordinate pair corresponding to a pair of conserved physical quantities; and the Rayleigh dissipation function is configured to characterize the energy loss in the grid-type energy storage system, including switching conduction losses, virtual synchronous machine damping losses, thermal conduction losses, and battery internal resistance losses.

[0041] In a preferred embodiment of this application, a two-stage parameter acquisition strategy combining offline initial identification and online fine-tuning is adopted to determine the kinetic energy metric tensor g. In the offline initial identification stage, a pre-designed specific mode current disturbance is injected into the grid-connected energy storage system in its off-grid state, and the system's transient energy response is measured. Based on the correspondence between the disturbance excitation and the energy response, the matrix elements of the kinetic energy metric tensor are determined. It can be preliminarily identified that its value is proportional to the component of the system transient energy response related to mode j caused by mode i. During the online fine-tuning phase, after grid connection, utilizing the inherent background disturbances of the power grid (such as small load fluctuations and harmonics), a recursive least squares estimation algorithm with physical structure preservation constraints (such as symplectic structure constraints and passive constraints) is used to... Real-time, recursive fine-tuning is performed to ensure the continuous accuracy of its characterization of the inertial coupling strength between state variables under all operating conditions.

[0042] In a preferred embodiment of this application, the core function of the canonical symplectic structure is to uniquely determine canonical coordinate pairs with definite physical meaning (each pair corresponds to a pair of physical conservation quantities), and the specific coordinate pairs are as follows:

[0043] The fast-scale canonical coordinate pairs for grid-type energy storage systems are as follows:

[0044] ( , = (inductance flux, inductance current);

[0045] ( , ) = (capacitor charge, capacitor voltage).

[0046] The mesoscale canonical coordinate pairs for grid-type energy storage systems are as follows:

[0047] ( , ) = (virtual rotor angle, angular velocity);

[0048] ( , = (Voltage loop integral, voltage amplitude).

[0049] The slow-scale canonical coordinate pairs for grid-type energy storage systems are as follows:

[0050] ( , ) = (Battery accumulated charge, battery terminal voltage);

[0051] ( , = (Accumulated heat, thermodynamic potential).

[0052] Preferably, the Poisson bracket relationship is verified. And with other parentheses set to zero, the physical determinism of the regular coordinate pairs is confirmed. For generalized coordinates, For generalized momentum, Let Kronecker function be used.

[0053] In a preferred embodiment of this application, the mathematical expression for the Rayleigh dissipation function is as follows:

[0054]

[0055] in, Represents the Rayleigh dissipation function. Indicates the on-resistance of the switch. This represents the rate of change of charge in a capacitor. This represents the damping coefficient of the virtual synchronizer. This represents the rate of change of the virtual rotor angular velocity. Indicates the thermal conductivity coefficient of the battery. Indicates the rate of change of battery pack temperature. This represents the total internal resistance of the battery, including ohmic internal resistance and polarization internal resistance. This indicates the rate of change of the battery's state of charge.

[0056] The above scheme accurately classifies the full-state variables of the grid-type energy storage system according to the dominant dynamic time scale, and combines continuous processing to ensure the smoothness of the state space. It constructs a three-layer nested differential manifold and defines geometric structures such as kinetic energy metric tensor, canonical symplectic structure, and Rayleigh dissipation function. This not only provides a geometric state space carrier that fits the physical essence for subsequent unified modeling, but also quantifies the time ratio relationship at different scales, the inertial coupling strength between state variables, physical conservation characteristics, and various energy losses, thus ensuring the physical consistency and integrity of the modeling from the root.

[0057] S200: Based on the geometric state space, a generalized Lagrange unified dynamic equation for a grid-type energy storage system is established. By introducing scale parameters, dynamics at multiple time scales are explicitly separated, and the coupling relationship between dynamics at different time scales is quantified using Lie bracket operations.

[0058] In the embodiments of this application, the kinetic energy T and potential energy V of the grid-type energy storage system are defined based on the geometric state space of the grid-type energy storage system.

[0059]

[0060] in, and Let be the first derivative of each state variable of the grid-type energy storage system with respect to time, i.e., the rate of change of state; i and j are tensor indices.

[0061]

[0062] in, Electromagnetic potential energy refers to the energy stored in the conservative electromagnetic field within a grid-type energy storage system. Represents synchronous potential energy, a virtual potential field introduced by the virtual synchronizer control, simulating the synchronization capability of a synchronous generator; It represents chemical potential energy, the chemical energy stored by the electrochemical reactions inside the battery.

[0063] The generalized Lagrange unified dynamic equations for grid-type energy storage systems are derived from the Rayleigh dissipation function of the grid-type energy storage system as follows:

[0064]

[0065] in, L represents the Lagrangian quantity of a grid-type energy storage system. For the partial derivatives of the Lagrange quantities with respect to the state variables of the grid-type energy storage system, For the partial derivatives of the Lagrange with respect to the state variables of the grid-type energy storage system, Let be the partial derivative of the Rayleigh dissipation function with respect to the state variables of the grid-type energy storage system. This refers to the external generalized force acting on a grid-type energy storage system.

[0066] In the embodiments of this application, explicit separation of multi-timescale dynamics is achieved by introducing scale parameters, including: using the first scale separation parameter and the second scale separation parameter as timescale transformation factors for the corresponding dynamic terms, and introducing them into the generalized Lagrange unified dynamic equation; multiplying the first scale separation parameter by all generalized inertial terms in the equation related to mesoscale dynamics, and multiplying the second scale separation parameter by all generalized inertial terms in the equation related to slow-scale dynamics, to obtain the scale-reconstructed generalized Lagrange unified dynamic equation; and reconstructing the scale-reconstructed generalized Lagrange unified dynamic equation into three mutually coupled subsystems to achieve explicit separation of multi-timescale dynamics.

[0067] The generalized Lagrange unified dynamic equations for scale reorganization are as follows:

[0068]

[0069]

[0070]

[0071] in, , , Let be the tensors on fast, medium, and slow-scale submanifolds, respectively. , , These represent the generalized forces at fast, medium, and slow scales, respectively, and u is the control input vector. , , These are the second-order time derivatives of the state vectors at fast, medium, and slow scales, respectively. It is the first derivative of the complete state vector of the grid-type energy storage system.

[0072] In the embodiments of this application, the three mutually coupled subsystems are a first timescale subsystem, a second timescale subsystem, and a third timescale subsystem. The first timescale subsystem characterizes the electromagnetic transient process dominated by the natural time constant of the grid-type energy storage system, corresponding to the dynamic characteristics of the fast-scale state set. The second timescale subsystem characterizes the electromechanical transient process and control dynamic process dominated by the time constant slowed down by the first-scale separation parameter, corresponding to the dynamic characteristics of the mesoscale state set. The third timescale subsystem characterizes the energy dispatching process and aging evolution process dominated by the time constant slowed down by the second-scale separation parameter, corresponding to the dynamic characteristics of the slow-scale state set.

[0073] In the embodiments of this application, the coupling relationship between dynamics at different time scales is quantified using Lie bracket operations, including: defining a fast-scale vector field, a medium-scale vector field, and a slow-scale vector field based on the generalized Lagrange unified dynamic equations, respectively, for fast-scale state sets, medium-scale state sets, and slow-scale state sets; calculating a first Lie bracket between the fast-scale vector field and the medium-scale vector field, a second Lie bracket between the medium-scale vector field and the slow-scale vector field, and a third Lie bracket between the fast-scale vector field and the slow-scale vector field; quantifying the nonlinear coupling strength between fast-scale dynamics and medium-scale dynamics, between medium-scale dynamics and slow-scale dynamics, and between fast-scale dynamics and slow-scale dynamics, respectively, according to the magnitudes of the first, second, and third Lie brackets; and mapping each quantized nonlinear coupling strength to engineering parameters including coupling gain and coupling delay.

[0074] Three types of reference gains for grid-type energy storage systems are defined as reference benchmarks for coupling strength mapping. The reference gains are as follows:

[0075]

[0076]

[0077]

[0078] in, This represents the average power of a grid-type energy storage system. This represents the harmonic voltage of a grid-type energy storage system. The value represents the rate of change of the state of charge of the battery, and P represents the power of the grid-type energy storage system. This indicates the average temperature of the battery. This represents the harmonic losses of a grid-type energy storage system. The reference gain represents the sensitivity of harmonic voltage to average power. This indicates the sensitivity of power to the rate of change of SOC. This indicates the sensitivity of harmonic loss to average temperature rise.

[0079] Furthermore, the normalized coupling gain of the grid-type energy storage system is calculated using the following formula:

[0080]

[0081] in, This represents the normalized gain of cross-scale dynamic coupling. Indicates the reference gain parameter. This represents the magnitude of the result of the Lie bracket operation on two vector fields at different time scales. This represents the maximum value of the Ligand bracket modulus under all operating conditions of a grid-type energy storage system.

[0082] Furthermore, cross-correlation analysis was used to determine the three types of coupling delays in the grid-type energy storage system, while ensuring that each coupling delay satisfies physical causality constraints. The determined coupling delays are as follows:

[0083]

[0084] in, This represents the time delay parameters corresponding to various types of coupling. , These are the typical time constants for the two scales in the corresponding coupling relationship.

[0085] The above scheme derives the generalized Lagrange unified dynamic equation based on the geometric state space, achieves explicit separation of dynamics across multiple time scales through scale parameters, and accurately quantifies the cross-scale nonlinear coupling strength by using Lie bracket operations and maps it to engineering coupling gain and time delay parameters. This not only establishes a unified dynamic expression that spans all scales, but also provides a quantitative basis for the cross-scale linkage of subsequent parameterized equivalent models.

[0086] S300: Based on the unified dynamic equation and the quantized coupling relationship, construct a parameterized equivalent model and perform collaborative optimization on the parameterized equivalent model.

[0087] In the embodiments of this application, a parameterized equivalent model is constructed and collaboratively optimized, including: constructing a parameterized equivalent model, which includes a first equivalent module (fast dynamic equivalent module), a second equivalent module (medium dynamic equivalent module), a third equivalent module (slow dynamic equivalent module), and a cross-scale coupling module; the first equivalent module corresponds to fast-scale dynamics, and includes a controlled voltage source and a frequency-varying impedance network. The amplitude and phase of the controlled voltage source are driven in real time by the fast-scale state set (filter inductor current, capacitor voltage, and switching duty cycle function). The frequency-varying impedance network uses piecewise linear fitting to characterize the impedance characteristics at different frequencies, and the core parameters include equivalent inductance, equivalent capacitance, and equivalent switching resistance. The second equivalent module corresponds to medium-scale dynamics, and is an extended virtual synchronous machine model, which extends the phase-locked loop submodule and power control loop submodule based on the standard virtual synchronous machine model. The third equivalent module corresponds to slow-scale dynamics. The third equivalent module includes time-varying parameters and a time-varying parameter battery thermal coupling model that integrates aging prediction. The time-varying parameters cover the equivalent parameters corresponding to the battery state of charge and average temperature. The battery thermal coupling model integrates electrochemical-thermal-aging coupling equations. The core state is the battery's internal health state, the input is the scheduling command, and the output is the battery terminal voltage and real-time aging rate. The cross-scale coupling module is built based on quantized coupling gain and coupling delay, and adopts a dual-unit structure of "gain adjustment + delay compensation". It establishes interactive channels for fast-medium, medium-slow, and fast-slow scales respectively: The fast-medium scale channel inputs the virtual synchronous machine angular velocity signal output from the second equivalent module into the first equivalent module after fast-medium scale coupling gain adjustment and fast-medium scale coupling delay compensation; The medium-slow scale channel inputs the battery state of charge signal output from the third equivalent module into the second equivalent module after medium-slow scale coupling gain adjustment and medium-slow scale coupling delay compensation; The fast-slow scale channel inputs the harmonic current signal output from the first equivalent module into the third equivalent module after fast-slow scale coupling gain adjustment and fast-slow scale coupling delay compensation, thereby realizing multi-scale dynamic coordinated linkage.

[0088] In the embodiments of this application, the parameters of the parameterized equivalent model are initially identified step by step to obtain the initial values ​​of the parameters of each module of the parameterized equivalent model; the coupling gain and coupling delay of the cross-scale coupling module are used as coupling constraints, and the parameters of all modules are optimized collaboratively starting from the initial values ​​of the parameters; the collaborative optimization is to minimize the output error of the parameterized equivalent model under the condition of satisfying the coupling constraints.

[0089] In a preferred embodiment of this application, the parameters of the first equivalent module are initially identified by an off-grid wideband sweep frequency test. A sweep frequency voltage signal is injected into the grid-type energy storage system, and fast-scale port voltage and current data are collected. The equivalent inductance, equivalent capacitance, and equivalent resistance parameters of the frequency-variable impedance network are identified by a vector fitting method. The gain parameters of the controlled voltage source are calibrated by combining the voltage-duty cycle relationship under static conditions, and the initial values ​​of all parameters of the first equivalent module are obtained.

[0090] The initial identification of the second equivalent module parameters was carried out using a grid-connected power step test (power step amplitude ±10% of rated power). Data on the virtual synchronous machine angular velocity, output active and reactive power, and phase-locked loop output angle were collected. Based on the model reference adaptive algorithm, the core mechanical parameters such as equivalent moment of inertia and damping coefficient were first identified. Then, the initial values ​​of the proportional parameters, integral parameters, active droop coefficient, and reactive droop coefficient of the phase-locked loop and power control loop were obtained through the proportional-integral-derivative parameter tuning method.

[0091] The initial identification of the third equivalent module parameters adopts a battery full-condition test combination scheme, including hybrid pulse power characteristic test (identifying the battery equivalent resistance and capacitance), full-temperature static charge and discharge test (minus 20 degrees Celsius to 60 degrees Celsius, identifying temperature-related time-varying parameters), and accelerated aging test (1,000 cycles, identifying aging prediction model parameters). The test data is fitted by the least squares method to obtain the initial values ​​of the equivalent circuit parameters, thermal conductivity coefficient, and aging rate coefficient of the battery thermal coupling model.

[0092] The initial parameters of the cross-scale coupling module are initially identified by directly reusing the initial values ​​of coupling gain obtained through Lie bracket quantization and mapping, combined with the initial values ​​of coupling delay obtained through cross-correlation analysis, as the initial parameters of the cross-scale coupling module.

[0093] In a preferred embodiment of this application, the Levenberg-Marquardt algorithm (LM) is used as the core optimization algorithm, which can efficiently solve multi-parameter coupled optimization problems.

[0094] The first stage of optimization involves fixing the parameters of the cross-scale coupling module, optimizing the independent parameters of the first, second, and third equal-value modules, reducing the error within a single module, and obtaining the optimized parameters for each module.

[0095] The second stage of optimization involves releasing the constraints on the parameters of the coupled modules, incorporating all module parameters into the optimization, and performing global collaborative optimization with the goal of minimizing the overall output error. During the process, the satisfaction of the coupling constraints is verified in real time.

[0096] Set convergence criteria: when the relative change of the objective function value between two adjacent iterations is <1e-6 and the change of all parameters is <1e-4, stop optimization and output the optimal parameter set.

[0097] Further optimization involves verifying the effectiveness of the optimized parametric equivalent model through multi-condition testing. If the error does not meet the requirements, closed-loop correction is performed to ensure that the model accuracy meets engineering application standards.

[0098] The above solution constructs equivalent modules and cross-scale coupling modules corresponding to fast, medium and slow scale dynamics. It obtains reasonable initial parameters through scenario-by-scenario step-by-step initial identification, and then minimizes the model output error through module-independent parameter optimization and global collaborative optimization. This enables the equivalent model to accurately match the core dynamic characteristics of each scale, and achieve multi-scale collaborative linkage through cross-scale channels with gain adjustment and time delay compensation, thereby improving engineering adaptability and accuracy.

[0099] S400: The parameterized equivalent model after collaborative optimization is verified under all operating conditions, and a hierarchical parallel adaptive update mechanism is deployed to enable the parameterized equivalent model to continuously evolve throughout its life cycle, thereby achieving unified modeling and physical consistency equivalence of grid-type energy storage systems across multiple time scales.

[0100] In the embodiments of this application, the constrained feedforward update layer in the hierarchical parallel adaptive update mechanism updates the parameters of the parameterized equivalent model online using a constrained recursive estimation algorithm during the grid-connected operation of the grid-type energy storage system.

[0101] In a preferred embodiment of this application, the parameterized equivalent model after collaborative optimization is first validated under all operating conditions. The validation covers typical scenarios including normal rated operation of the grid-connected energy storage system, power step regulation, extreme faults such as grid voltage drops or frequency fluctuations, and battery aging throughout its entire lifecycle. By comparing the model output with measured system data, the dynamic response accuracy and physical consistency of the model under different operating conditions are verified. After successful validation, a hierarchical parallel adaptive update mechanism is deployed to ensure the model's effectiveness throughout its entire lifecycle. The core of this mechanism includes a constrained feedforward update layer. This update layer can collect system operation data in real time during the grid-connected operation of the grid-connected energy storage system and use a constrained recursive estimation algorithm to update the key parameters of the parameterized equivalent model online. During the update process, core conditions such as physical boundary constraints and cross-scale coupling constraints are strictly followed to ensure the rationality and safety of parameter updates. Through the coordinated implementation of the above-mentioned full-condition verification and hierarchical parallel adaptive update, the parameterized equivalent model can continuously evolve throughout its entire life cycle as the system's operating state changes and components age. Ultimately, it stably achieves the multi-timescale unified modeling and physical consistency equivalence goals of the grid-type energy storage system, providing reliable model support for the efficient and stable operation of the system.

[0102] The above scheme ensures the universality and reliability of the optimized model through full-condition verification covering scenarios such as normal operation, extreme failure, and full life cycle aging. The hierarchical parallel adaptive update mechanism, with the help of a constrained online recursive estimation algorithm, allows the parameterized equivalent model to continuously evolve with changes in the operating status of the grid-type energy storage system and factors such as component aging. Ultimately, it achieves unified modeling and physical consistency equivalence across multiple time scales throughout the life cycle of the grid-type energy storage system, providing solid model support for the efficient and stable operation of the system.

[0103] This invention also provides a unified modeling and equivalence system for multi-timescale grid-type energy storage systems, used to implement a unified modeling and equivalence method for multi-timescale grid-type energy storage systems. The system includes a control module, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program to implement a unified modeling and equivalence method for multi-timescale grid-type energy storage systems.

[0104] This invention provides a storage medium storing a program that, when executed by a processor, implements a unified modeling and equivalence method for multi-timescale energy storage systems.

[0105] This invention provides a processor for running a program, wherein the program executes a unified modeling and equivalence method for multi-timescale energy storage systems during runtime.

[0106] This invention provides a device including a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, it implements a unified modeling and equivalence method for multi-timescale energy storage systems. The device described herein can be a server, PC, tablet, mobile phone, etc.

[0107] This application also provides a computer program product that, when executed on a data processing device, is suitable for performing a unified modeling and equivalence method for multi-timescale energy storage systems.

[0108] Those skilled in the art will understand that embodiments of this application can provide methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0109] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0110] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0111] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0112] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0113] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0114] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0115] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0116] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A unified modeling and equivalence method for multi-timescale energy storage systems, characterized in that, include: The full-state variables of the grid-type energy storage system are classified according to the time scale of the dominant dynamic, and a three-layer nested differential manifold is constructed based on the time scale classification to obtain the geometric state space. Based on the geometric state space, the generalized Lagrange unified dynamic equation of the grid-type energy storage system is established. The dynamics of multiple time scales are explicitly separated by introducing scale parameters, and the coupling relationship between dynamics of different time scales is quantified by using Lie bracket operations. Based on the unified dynamic equation and the quantized coupling relationship, a parameterized equivalent model is constructed, and the parameterized equivalent model is collaboratively optimized. The parameterized equivalent model after collaborative optimization is verified under all operating conditions, and a hierarchical parallel adaptive update mechanism is deployed to enable the parameterized equivalent model to continuously evolve throughout its entire life cycle, thereby achieving unified modeling and physical consistency equivalence of the grid-type energy storage system across multiple time scales.

2. The multi-timescale unified modeling and equivalence method for grid-type energy storage systems according to claim 1, characterized in that, The total state variables are classified into fast-scale state sets, medium-scale state sets, and slow-scale state sets according to the time scale of the dominant dynamic.

3. The multi-timescale unified modeling and equivalence method for grid-type energy storage systems according to claim 2, characterized in that, The construction of a three-layer nested differential manifold based on the time scale classification yields a geometric state space, including: Construct corresponding submanifolds for the fast-scale state set, the medium-scale state set, and the slow-scale state set, respectively; The submanifolds are combined into smooth product manifolds to form a three-layer nested differential manifold; The ratio of the time constants of fast-scale dynamics to mesoscale dynamics is defined as the first-scale separation parameter, and the ratio of the time constants of mesoscale dynamics to slow-scale dynamics is defined as the second-scale separation parameter.

4. The multi-timescale unified modeling and equivalence method for grid-type energy storage systems according to claim 1, characterized in that, The construction of a three-layer nested differential manifold based on the time scale classification to obtain the geometric state space also includes: Define the geometric structure in the geometric state space; The geometric structure includes a kinetic energy metric tensor, a canonical symplectic structure, and a Rayleigh dissipation function, wherein... The kinetic energy metric tensor is configured to quantitatively characterize the inertial coupling strength between state variables through its matrix elements; The canonical symplectic structure is configured to uniquely determine multiple canonical coordinate pairs, each of which corresponds to a pair of physical conservation quantities. The Rayleigh dissipation function is configured to characterize the energy loss in the grid-type energy storage system, which includes switching conduction loss, virtual synchro damping loss, thermal conduction loss, and battery internal resistance loss.

5. The multi-timescale unified modeling and equivalence method for grid-type energy storage systems according to claim 3, characterized in that, The explicit separation of dynamics across multiple time scales by introducing a scale parameter includes: The first scale separation parameter and the second scale separation parameter are used as time scale transformation factors for the corresponding dynamic terms and introduced into the generalized Lagrange unified dynamic equation. Multiply the first scale separation parameter by all the generalized inertial terms in the equation that are related to mesoscale dynamics, and multiply the second scale separation parameter by all the generalized inertial terms in the equation that are related to slow-scale dynamics, to obtain the generalized Lagrange unified dynamic equation for scale reorganization. The generalized Lagrange unified dynamic equations reconstructed by the scale are reconstructed into three mutually coupled subsystems to achieve explicit separation of dynamics across multiple time scales. The three mutually coupled subsystems are the first time-scale subsystem, the second time-scale subsystem, and the third time-scale subsystem. The first timescale subsystem characterizes the electromagnetic transient processes dominated by the natural time constant of the grid-type energy storage system. The second time-scale subsystem characterizes the electromechanical transient processes and control dynamic processes dominated by time constants slowed down by the separation parameters of the first scale. The third timescale subsystem characterizes the energy scheduling process and aging evolution process dominated by the time constant slowed down by the second-scale separation parameters.

6. The multi-timescale unified modeling and equivalence method for grid-type energy storage systems according to claim 2, characterized in that, The method of quantifying the coupling relationship between dynamics at different time scales using Lie bracket operations includes: Based on the generalized Lagrange unified dynamic equations, fast-scale vector fields, mesoscale vector fields, and slow-scale vector fields are defined for fast-scale state sets, mesoscale state sets, and slow-scale state sets, respectively. Calculate the first Lie bracket between the fast-scale vector field and the mesoscale vector field, the second Lie bracket between the mesoscale vector field and the slow-scale vector field, and the third Lie bracket between the fast-scale vector field and the slow-scale vector field; Based on the modulus of the first, second, and third Lie brackets, the nonlinear coupling strength between fast-scale dynamics and mesoscale dynamics, between mesoscale dynamics and slow-scale dynamics, and between fast-scale dynamics and slow-scale dynamics are quantified respectively. The quantified nonlinear coupling strengths are mapped to engineering parameters that include coupling gain and coupling delay.

7. The multi-timescale unified modeling and equivalence method for grid-type energy storage systems according to claim 6, characterized in that, The construction of the parameterized isometry model and the collaborative optimization of the parameterized isometry model include: A parameterized isometry model is constructed, which includes a first isometry module, a second isometry module, a third isometry module, and a cross-scale coupling module; The first equivalent module corresponds to fast-scale dynamics, and the first equivalent module includes a controlled voltage source and a frequency-varying impedance network; The second equivalent module corresponds to the mesoscale dynamics, and the second equivalent module is an extended virtual synchronizer model; The third equivalent module corresponds to the slow-scale dynamics, and the third equivalent module includes time-varying parameters and a time-varying parameter battery thermal coupling model that integrates aging prediction.

8. The multi-timescale unified modeling and equivalence method for grid-type energy storage systems according to claim 7, characterized in that, The construction of the parameterized isometry model and the collaborative optimization of the parameterized isometry model further include: The parameters of the parameterized equivalent model are initially identified step by step to obtain the initial parameter values ​​of each module of the parameterized equivalent model; Using the coupling gain and coupling delay of the cross-scale coupling module as coupling constraints, and starting from the initial values ​​of the parameters, the parameters of all modules are optimized collaboratively. The collaborative optimization is to minimize the output error of the parameterized equivalent model while satisfying the coupling constraints.

9. The multi-timescale unified modeling and equivalence method for grid-type energy storage systems according to claim 1, characterized in that, The constrained feedforward update layer in the hierarchical parallel adaptive update mechanism updates the parameters of the parameterized equivalent model online during the grid-connected operation of the grid-type energy storage system using a constrained recursive estimation algorithm.

10. A unified modeling and equivalent system for multi-timescale energy storage systems, characterized in that, The system includes a control module, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program to implement the multi-timescale unified modeling and equivalence method for grid-type energy storage systems according to any one of claims 1-9.

Images