Micro-grid-oriented photovoltaic energy storage device comprehensive operation control method and system

Abstract

The present application relates to the technical field of energy control, and especially relates to a micro-grid-oriented photovoltaic storage device comprehensive operation control method and system.The method comprises the following steps: obtaining current and voltage waveform data at both ends of a filter inductor of the photovoltaic storage device; performing frequency domain sweep on the current and voltage waveform data to generate filter inductor complex impedance data; mapping the filter inductor complex impedance data to obtain a local saturation starting point and a nonlinear section of a magnetic core, and generating saturation characteristic data; connecting a bidirectional DC bias winding in parallel to the filter inductor, and adjusting the saturation characteristic data to generate magnetic core working point adjustment data.The present application solves the problems of unstable filter performance, fixed modulation parameters and insufficient system cooperation of traditional photovoltaic storage devices by accurately identifying the nonlinear characteristics of the filter inductor, dynamically adjusting the working point of the magnetic core, adaptively optimizing the PWM modulation parameters and micro-grid collaborative control.

Description

Technical Field

[0001] This invention relates to the field of energy control technology, and in particular to a method and system for integrated operation control of photovoltaic and energy storage devices for microgrids. Background Technology

[0002] Traditional power grids have limited capacity for integrating and regulating distributed energy resources, resulting in low operating efficiency of photovoltaic (PV) and energy storage (ESS) devices and failing to meet the dual demands of microgrids for stability and economy. Early PV and ESS device operation and control methods relied heavily on single devices or centralized control, lacking adaptability to system diversity and dynamic changes, easily leading to power fluctuations and wasted energy storage resources. With the rise of smart grids and IoT technologies, distributed control strategies based on real-time data acquisition and edge computing have been proposed, effectively improving the collaborative scheduling capabilities and response speed of PV and ESS devices. Simultaneously, optimization algorithms based on big data and artificial intelligence are being applied to the operation and management of PV and ESS systems, driving control methods towards intelligence and precision. However, traditional PV and ESS device filter inductors often suffer from local core saturation and nonlinearity, leading to unstable filtering effects and affecting the system's harmonic suppression capability. Furthermore, the PWM carrier slope and dead zone are often designed with fixed values, making it difficult to adapt to filter nonlinearity and system dynamic changes, resulting in increased switching losses and insufficient harmonic suppression. Summary of the Invention

[0003] Therefore, it is necessary to provide a comprehensive operation control method and system for photovoltaic and energy storage devices in microgrids to solve at least one of the above-mentioned technical problems.

[0004] To achieve the above objectives, a comprehensive operation and control method for photovoltaic and energy storage devices in microgrids is provided, the method comprising the following steps:

[0005] Step S1: Obtain the current and voltage waveform data across the filter inductor of the photovoltaic energy storage device; perform frequency domain sweep on the current and voltage waveform data to generate complex impedance data of the filter inductor; based on the complex impedance data of the filter inductor, map the local saturation start point and nonlinear segment of the magnetic core to generate saturation characteristic data;

[0006] Step S2: Connect a bidirectional DC bias winding in parallel next to the filter inductor, and adjust the bias current based on the saturation characteristic data to generate core operating point adjustment data; split the filter into two groups, main and auxiliary, and dynamically switch them in parallel / series under different operating conditions using solid-state relays, and configure the filter impedance based on the saturation characteristic data to generate filter impedance configuration data.

[0007] Step S3: Adaptively adjust the PWM carrier slope and dead zone of the photovoltaic energy storage device using the magnetic core operating point adjustment data and filter impedance configuration data to obtain PWM modulation parameter optimization data; perform virtual inductance / resistance compensation superposition on the PWM modulation parameter optimization data to generate adaptive compensation control data;

[0008] Step S4: Generate control commands from the adaptive compensation control data, and input the generated control commands into the preset microgrid for command-coordinated control, so as to perform integrated operation control optimization of photovoltaic and energy storage equipment for microgrid.

[0009] This invention acquires the current and voltage waveform data across the filter inductor of a photovoltaic energy storage device and performs frequency domain sweep analysis to accurately obtain the complex impedance characteristics of the filter inductor. This allows for the mapping of the local saturation start point and nonlinear segment of the magnetic core, achieving precise modeling of the magnetic core saturation characteristics. Furthermore, a bidirectional DC bias winding is connected in parallel with the filter inductor, and the magnetic core saturation characteristics are dynamically adjusted, enabling controllable adjustment of the magnetic core operating point. By dividing the filter inductor into two groups, main and auxiliary, and using solid-state relays to achieve intelligent parallel / series switching under different operating conditions, the equivalent impedance of the filter can be flexibly configured under complex operating conditions, enhancing the electromagnetic compatibility and stability of the system. Based on this, by combining the magnetic core operating point adjustment data and the filter impedance configuration data, the PWM carrier slope and dead zone are adaptively adjusted. This not only optimizes the PWM modulation strategy but also effectively suppresses voltage and current distortion through the superposition of virtual inductance and virtual resistance compensation. The resulting adaptive compensation control command can work collaboratively with other control units in the microgrid, significantly improving the operational stability, response flexibility, and overall energy efficiency of the photovoltaic energy storage device in the microgrid. Therefore, this invention solves the problems of unstable filtering performance, fixed modulation parameters, and insufficient system coordination in traditional photovoltaic energy storage equipment by accurately identifying the nonlinear characteristics of the filter inductor, dynamically adjusting the operating point of the magnetic core, adaptively optimizing PWM modulation parameters, and coordinating with the microgrid.

[0010] Preferably, step S1 includes the following steps:

[0011] Step S11: Use a current probe and a differential voltage probe to synchronously acquire transient waveforms on the input and output sides of the filter inductor to obtain current and voltage waveform data across the filter inductor of the photovoltaic energy storage device.

[0012] Step S12: Perform a fast Fourier transform on the current and voltage waveform data to generate frequency domain swept frequency response spectrum data; calculate the amplitude-phase ratio based on the frequency domain swept frequency response spectrum data to generate the complex impedance data of the filter inductor;

[0013] Step S13: Perform inverse admittance phasor transformation analysis on the complex impedance data of the filter inductor, extract the equivalent permeability nonlinear inflection point, and generate the starting point data of local saturation of the magnetic core; based on the starting point data of local saturation of the magnetic core, perform segment trend slope segmentation on the complex impedance data of the filter inductor, identify the nonlinear gain interval, and generate the nonlinear segment data of the magnetic core.

[0014] Step S14: Merge the local saturation start point data of the magnetic core with the nonlinear section data of the magnetic core to generate saturation characteristic data.

[0015] This invention employs high-precision synchronous sampling of both ends of a filter inductor using current and differential voltage probes, effectively capturing transient response characteristics during the operation of photovoltaic energy storage devices and improving the integrity and accuracy of waveform data. Utilizing Fast Fourier Transform (FFT) technology, frequency domain analysis is performed on the acquired current and voltage waveforms to construct a frequency-domain swept response spectrum, thereby achieving precise modeling of the impedance behavior of the filter inductor at different frequencies. The complex impedance data calculated based on the frequency domain amplitude and phase ratio accurately reflects the electromagnetic characteristics of the magnetic core under different operating conditions. Further analysis of the complex impedance data through inverse admittance phasor transformation accurately extracts the nonlinear inflection point of the equivalent permeability, thus identifying the starting point of the magnetic core entering local saturation. Furthermore, the slope segmentation method is used to classify the trends of complex impedance segments, clarifying the distribution boundary of the nonlinear gain region and avoiding the ambiguity and error accumulation in nonlinear segment modeling using traditional methods. Finally, the information on the local saturation starting point and nonlinear segments is integrated to form high-resolution magnetic core saturation characteristic data, providing a scientific and quantifiable foundation for subsequent magnetic core operating point adjustment and filter dynamic configuration.

[0016] Preferably, in step S13, performing inverse admittance phasor transformation analysis on the complex impedance data of the filter inductor to extract the equivalent permeability nonlinear inflection point includes:

[0017] The complex impedance data of the filter inductor is conjugate normalized to generate unit amplitude admittance spectrum data.

[0018] The imaginary admittance curve is extracted from the unit amplitude admittance spectrum data to generate the inductive admittance change trajectory;

[0019] The slope change rate at local extreme points of the inductive admittance change trajectory is analyzed to generate a first-order slope derivative curve.

[0020] Identify abrupt slope changes in the first-order slope derivative curve, extract the inflection frequency points of magnetic permeability drops, and generate data on abrupt changes in magnetic permeability.

[0021] By combining the rate of change of complex impedance phase angle corresponding to the abrupt change frequency, the phase lag characteristic is verified, the key frequency points that conform to the core saturation response are output, and the data of the local saturation start point of the core are generated.

[0022] This invention constructs a unit-amplitude admittance spectrum by performing conjugate normalization on the complex impedance data of the filter inductor, effectively eliminating the interference of dimensional and amplitude differences on nonlinear feature identification and providing a unified metric basis for subsequent feature analysis. It extracts the imaginary component of the admittance spectrum to form an inductive admittance change trajectory, which can realistically reflect the inductive reactance response behavior of the magnetic core at different frequencies. Furthermore, by analyzing the slope change rate at local extreme points of the admittance trajectory, a first-order slope derivative curve is constructed, achieving a quantitative characterization of the abrupt change trend of inductive admittance. The invention then identifies... The slope abrupt change point effectively locates the turning frequency of the core permeability drop, thereby extracting key permeability change abrupt change point data; combined with the complex impedance phase angle change rate at the abrupt change point, it further determines whether there is a significant phase lag characteristic, and confirms from the frequency response perspective that the core has entered a local saturation state; the final output core local saturation start point data not only has high resolution and high robustness, but also significantly outperforms traditional saturation identification methods based on steady state or empirical models, ensuring the accuracy and dynamic adaptability of the physical basis for subsequent core operating point adjustment and nonlinear compensation.

[0023] Preferably, step S2 includes the following steps:

[0024] Step S21: Connect a bidirectional DC bias winding next to the filter inductor, and set the initial bias value of the adjustable constant current source based on the local saturation start point data of the magnetic core to generate the initial excitation parameters of the bias winding.

[0025] Step S22: Analyze the bias current regulation response curve of the saturation characteristic data to generate the gain coefficient curve of the bias current to the change of magnetic permeability; based on the gain coefficient curve of the bias current to the change of magnetic permeability, dynamically set the bias winding control command and generate core operating point adjustment data.

[0026] Step S23: Split the filter inductor into a main filter branch and an auxiliary filter branch, and configure independent sampling channels to measure the voltage and current of the two branches respectively, generating filter branch operating status monitoring data;

[0027] Step S24: Obtain the operating load data of the main filter branch and the auxiliary filter branch; based on the operating load data, identify the load disturbance mode of the filter branch operating status monitoring data and generate dynamic switching threshold condition data.

[0028] Step S25: Configure solid-state relay control logic, perform series-parallel dynamic switching of main and auxiliary filter branches based on dynamic switching threshold condition data and solid-state relay control logic, and generate filter topology switching status data; perform filter impedance mapping analysis on topology switching status data and saturation characteristic data to generate filter impedance configuration data under the current topology.

[0029] This invention introduces a bidirectional DC bias winding next to the filter inductor and configures an adjustable constant current source based on the local saturation start point data of the magnetic core, achieving precise injection control of the bias current. This effectively expands the operating range of the magnetic core and enhances its linear control capability. Furthermore, by analyzing the bias current adjustment response curve of the saturation characteristic data, a gain coefficient curve of bias current-permeability change is constructed, enabling quantitative modeling of the nonlinear characteristics of the magnetic core and providing data support for the dynamic adjustable optimization of the core's operating point. Moreover, the filter inductor is structurally divided into main and auxiliary branches, and independent sampling channels are configured for real-time monitoring, effectively improving the observability and management granularity of the filter branch's operating status. By combining load data with disturbance pattern identification of filter branch monitoring data, switching threshold condition data can be dynamically generated to support the configuration of series-parallel topology switching control logic for solid-state relays. This mechanism can intelligently select the main and auxiliary branch combination method according to load fluctuations, thereby realizing topology adaptive adjustment of the filter in complex operating environments. Finally, by combining topology switching status and saturation characteristic data, filter impedance mapping analysis is performed to generate the optimal filter impedance configuration data under the current topology. This not only improves the consistency and stability of filter performance, but also provides physical support for dynamic matching of subsequent PWM modulation parameter optimization, significantly enhancing the power quality control capability of photovoltaic energy storage equipment.

[0030] Preferably, step S24, which involves identifying load disturbance patterns in the filter branch operating status monitoring data based on the load data, includes:

[0031] Perform periodic stability assessment on the load data and generate load period drift characteristic data;

[0032] Extract the power fluctuation threshold from the filter branch operating status monitoring data to obtain power fluctuation mutation index data;

[0033] Joint pattern classification is performed on load cycle drift characteristic data and power fluctuation mutation index data to generate disturbance pattern classification label data, which includes high-frequency impact type, intermittent oscillation type and continuous jitter type.

[0034] Mapping rules are set for the disturbance pattern classification label data to generate disturbance response priority data;

[0035] Configure event trigger thresholds based on disturbance response priority data, and generate dynamic switching threshold condition data.

[0036] This invention performs periodic stability discriminant analysis on load data to construct load periodic drift characteristic data, effectively capturing periodic instability caused by load fluctuations and providing preliminary criteria for filter response adjustment. Simultaneously, it extracts power fluctuation thresholds from filter branch operating status monitoring data to generate power fluctuation mutation index data, enabling precise perception of power disturbances under highly dynamic operating conditions. Furthermore, by jointly classifying the periodic drift characteristics and power fluctuation mutation index, it clarifies the dynamic characteristic categories of load disturbances, classifying them into typical operating conditions such as high-frequency impact, intermittent oscillation, and continuous jitter, effectively improving the decision-making targeting of filter topology switching. Based on the classification results, a disturbance response priority rule mapping mechanism is introduced to set priority processing levels for different types of disturbances, thereby avoiding blind switching logic and resource waste. Finally, by configuring event trigger thresholds through disturbance response priority, adaptively adjustable dynamic switching threshold condition data is generated, significantly improving the timeliness, selectivity, and stability of filter topology adjustment under multiple operating conditions, and enhancing the overall system's anti-interference capability and dynamic adaptability to power disturbances.

[0037] Preferably, joint pattern classification of load cycle drift characteristic data and power fluctuation abrupt change index data includes:

[0038] Sliding cross-spectral focusing analysis is performed on load cycle drift characteristic data to generate frequency perturbation coupling index data;

[0039] The perturbation edge acceleration features of the power fluctuation abrupt change index data are extracted to obtain the abrupt change response intensity data;

[0040] A two-factor perturbation feature matrix is ​​constructed based on frequency perturbation coupling index data and mutation response intensity data. Non-convex fuzzy clustering is then performed to generate a perturbation distribution label map.

[0041] The regional density gradient projection of the disturbance distribution label map is calculated, and the projection morphology is determined to obtain disturbance pattern classification label data. Specifically, the projection morphology is determined as follows: when the projection morphology is a cluster-oblique impact type, it is marked as a high-frequency impact type; when the projection morphology is a divergence-loop type, it is marked as an intermittent oscillation type; and when the projection morphology is an expansion-resonance type, it is marked as a continuous shaking type.

[0042] This invention generates frequency disturbance coupling index data reflecting the degree of frequency coupling of disturbances by performing sliding cross-spectral focusing analysis on load cycle drift characteristic data, effectively improving the recognition resolution of cycle instability characteristics and frequency disturbance coupling behavior. Simultaneously, it extracts disturbance edge acceleration features from power fluctuation mutation indicators to construct mutation response intensity data, which is used to accurately characterize the transient impact amplitude of load disturbances. Based on two types of key disturbance factors, a two-factor disturbance feature matrix is ​​constructed, and a non-convex fuzzy clustering algorithm is introduced for adaptive classification, enabling disturbance pattern recognition to take into account both boundary ambiguity and morphological complexity, overcoming the limitations of traditional clustering algorithms in handling nonlinear disturbances. The method overcomes the bottleneck of perturbation; the clustering results are calculated by projecting the regional density gradient of the perturbation distribution label map, which further compresses the dimensionality and strengthens the spatial distribution characteristics of the perturbation mode; through structural discrimination of the projection morphology, three typical perturbation modes can be effectively divided into three types: clustering-oblique impact type (high frequency impact type), splitting-loop type (intermittent oscillation type), and expansion-resonance type (continuous jitter type), achieving high-accuracy labeling and semantic association of perturbation types; the method has strong generalization ability and real-time adaptability under high complex load conditions, providing accurate and controllable perturbation identification results for subsequent filter topology switching, and significantly improving the intelligent identification and response efficiency of the photovoltaic storage system to external perturbations.

[0043] Preferably, step S3, which involves adaptively adjusting the PWM carrier slope and dead zone of the photovoltaic energy storage device using magnetic core operating point adjustment data and filter impedance configuration data, includes:

[0044] Identify the nonlinear saturation range of the magnetic core operating point adjustment data;

[0045] Perform frequency response matching analysis on the filter impedance configuration data to generate dynamic impedance characteristic data of the filter.

[0046] Based on the nonlinear saturation range, the dynamic impedance characteristic data of the filter are used to generate PWM carrier slope adaptive adjustment rules, and the carrier slope adjustment strategy data is obtained.

[0047] Based on carrier slope adjustment strategy data, dead zone correction is performed on the optical storage device to generate dead zone time correction parameter data.

[0048] The carrier slope adjustment strategy data and dead time correction parameter data are jointly mapped to generate PWM modulation parameter optimization data.

[0049] This invention proactively senses the local nonlinear response characteristics of the magnetic core by identifying the nonlinear saturation range in the core operating point adjustment data, avoiding the power quality degradation risk caused by traditional fixed modulation strategies when the core enters saturation. Combined with filter impedance configuration data, frequency response matching analysis is performed to construct dynamic impedance characteristic data of the filter reflecting the dynamic behavior of the filter under actual operating conditions, effectively revealing the influence channel of carrier modulation on the filter's energy response. Based on this, an adaptive PWM carrier slope adjustment rule is constructed for the nonlinear range characteristics, generating personalized carrier slope adjustment strategy data, which can accurately optimize the modulation slope according to the degree of core nonlinearity and filter response. Furthermore, dead-time correction is performed on the photovoltaic-storage equipment based on the carrier strategy data, generating dead-time correction parameters to effectively avoid system instability caused by switching device hysteresis and cross-conduction. Finally, by jointly mapping the carrier slope and dead-time correction parameters, optimized PWM modulation parameter data with nonlinear adaptive capabilities is generated, significantly improving the real-time performance and matching of the modulation strategy, and enhancing the switching accuracy, filtering efficiency, and overall stability of the photovoltaic-storage equipment under complex operating conditions, supporting the efficient, low-loss, and highly compatible operation of photovoltaic-storage systems in microgrids.

[0050] Preferably, dead-zone correction of the photovoltaic storage device based on carrier slope adjustment strategy data includes:

[0051] Model the switching delay caused by dead zone on the carrier slope adjustment strategy data to generate switching delay error modeling data;

[0052] The dead time correction amount is calculated based on the switching delay error model data, and the dead time correction amount data is generated.

[0053] The dead time width of the optical storage device is corrected using the dead time correction data, and the corrected dead time width is adjusted by feedback to generate dead time correction parameter data.

[0054] This invention models the switching delay caused by dead time in carrier slope adjustment strategy data, constructing a switching delay error model that accurately reflects the dynamic response offset of the switch during PWM modulation. This effectively overcomes the shortcomings of traditional dead time settings, which rely on empirical parameters and have coarse adjustments. Based on this model, the dead time correction amount is calculated, which can dynamically generate dead time correction data for the device conduction hysteresis characteristics under different carrier slope conditions, achieving more refined switching dead time control. Furthermore, the dead time width is adjusted using the correction amount, and a feedback adjustment mechanism is introduced. This mechanism can track the dead time adjustment effect in real time and dynamically converge to the optimal control range, thereby generating robust dead time correction parameter data. This mechanism not only improves the symmetry and effectiveness of PWM modulation and significantly reduces voltage distortion, bridge arm oscillation, and switching losses caused by unreasonable dead time settings, but also enhances the robustness and dynamic adaptability of the modulation strategy to high-frequency disturbances, laying a key foundation for achieving high-precision and high-reliability modulation control of optical storage devices.

[0055] Preferably, step S4 includes the following steps:

[0056] Step S41: Perform dynamic instruction encoding on the adaptive compensation control data to generate multi-dimensional control instruction set data;

[0057] Step S42: Identify instruction redundancy conflicts in the multidimensional control instruction set data, eliminate redundant and mutually exclusive instructions, and thus generate deduplication and optimization data for control instructions;

[0058] Step S43: Input the deduplication and optimization data of the control commands into the microgrid power routing module to perform regional load response mapping and generate microgrid command matching mapping data;

[0059] Step S44: Calculate the control timing consistency of the microgrid command matching mapping data to obtain the collaborative control scheduling sequence data; input the collaborative control scheduling sequence data to the preset microgrid control interface for command distribution and execution, generate operation feedback response data, and perform integrated operation control optimization operation for microgrid-oriented photovoltaic and energy storage equipment.

[0060] This invention constructs a multi-dimensional control instruction set data oriented towards multi-dimensional operational objectives by dynamically encoding adaptive compensation control data, realizing multi-level expression of photovoltaic and energy storage device control strategies in temporal, spatial, and functional dimensions. Furthermore, it identifies and eliminates redundant and mutually exclusive control instructions in the instruction set, generating optimized deduplicated control instruction data, significantly reducing the risk of instruction conflicts and resource waste, and improving the execution efficiency and consistency of the control system. The optimized instruction data is input to the microgrid's power routing module for regional load response mapping, generating microgrid instruction matching mapping data, achieving precise alignment between the control strategy and the actual operating state of the microgrid. Further, control timing consistency calculations are performed on the matching mapping data to generate a globally coordinated control scheduling sequence, ensuring seamless and conflict-free collaborative operation among multiple devices and nodes. Finally, by inputting the scheduling sequence to the microgrid control interface, real-time instruction distribution and execution are completed, and operational feedback response data is generated, establishing a closed-loop linkage mechanism between the photovoltaic and energy storage devices and the microgrid. Overall, this step effectively improves the dynamic optimization capability of control commands, the adaptability of microgrids, and the collaborative control capability, enabling efficient operation, flexible scheduling, and intelligent response of photovoltaic-storage systems in microgrids, and enhancing their system-level operational stability and optimization under complex load scenarios.

[0061] This specification provides a comprehensive operation and control system for photovoltaic and energy storage devices in microgrids, used to execute the aforementioned comprehensive operation and control method for photovoltaic and energy storage devices in microgrids. This comprehensive operation and control system for photovoltaic and energy storage devices in microgrids includes:

[0062] The data acquisition module is used to acquire the current and voltage waveform data across the filter inductor of the photovoltaic energy storage device; to perform frequency domain sweep on the current and voltage waveform data to generate complex impedance data of the filter inductor; and to map the local saturation start point and nonlinear segment of the magnetic core based on the complex impedance data of the filter inductor to generate saturation characteristic data.

[0063] The data intervention configuration module is used to connect a bidirectional DC bias winding in parallel next to the filter inductor, and to adjust the bias current based on the saturation characteristic data to generate core operating point adjustment data; it also splits the filter into two groups, main and auxiliary, which are dynamically switched in parallel / series by solid-state relays under different operating conditions, and configures the filter impedance based on the saturation characteristic data to generate filter impedance configuration data.

[0064] The control optimization module is used to adaptively adjust the PWM carrier slope and dead zone of the photovoltaic energy storage device by using magnetic core operating point adjustment data and filter impedance configuration data to obtain PWM modulation parameter optimization data; and to perform virtual inductance / resistance compensation superposition on the PWM modulation parameter optimization data to generate adaptive compensation control data.

[0065] The instruction distribution module is used to generate control instructions from adaptive compensation control data and input the generated control instructions into a preset microgrid for instruction-coordinated control, so as to perform integrated operation control optimization of photovoltaic and energy storage equipment for microgrid.

[0066] The beneficial effects of this invention lie in the fact that by setting up a data acquisition module, high-precision acquisition of the voltage and current waveforms across the filter inductor of the photovoltaic energy storage device is achieved. Furthermore, by using frequency domain sweep technology to extract the complex impedance characteristics of the filter inductor, the local saturation start point and nonlinear segment of the magnetic core are identified, effectively constructing a high-resolution saturation characteristic model reflecting the dynamic saturation behavior of the magnetic core, providing precise physical support for subsequent control strategies. The data intervention configuration module, by introducing an adjustable DC bias winding and a magnetic core operating point adjustment mechanism, enhances the controllability and linear range of the magnetic core. It also supports dynamic reconfiguration and series-parallel switching of the main and auxiliary branches of the filter inductor under multiple operating conditions. Combined with saturation characteristic data, intelligent configuration of the filter impedance is performed, significantly improving the filtering performance. The invention enhances the equivalent impedance adaptation capability and response sensitivity of the device under complex operating conditions. A control optimization module integrates core operating point adjustment data and filter impedance configuration data to achieve dynamic adaptive adjustment of the PWM carrier slope and dead zone. Furthermore, virtual inductance / resistance compensation is superimposed, enabling fine-grained tuning of control parameters. This effectively reduces switching losses, harmonic distortion, and system response lag, improving the real-time performance and stability of the modulation strategy. The instruction distribution module supports intelligent encoding, conflict avoidance, and microgrid scenario matching of adaptive compensation control data, thereby achieving coordinated distribution and closed-loop feedback of control instructions. This significantly improves the control coordination, operating efficiency, and intelligent response level of the photovoltaic-storage equipment in the microgrid. Therefore, this invention solves the problems of unstable filtering performance, fixed modulation parameters, and insufficient system coordination in traditional photovoltaic-storage equipment by accurately identifying the nonlinear characteristics of the filter inductor, dynamically adjusting the core operating point, adaptively optimizing PWM modulation parameters, and implementing microgrid coordinated control. Attached Figure Description

[0067] Figure 1 A flowchart illustrating the steps of a comprehensive operation and control method for photovoltaic and energy storage devices in a microgrid.

[0068] Figure 2 for Figure 1 A detailed flowchart illustrating the implementation steps of step S1.

[0069] Figure 3 for Figure 1 A detailed flowchart illustrating the implementation steps of step S4.

[0070] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0071] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0072] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0073] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0074] To achieve the above objectives, please refer to Figures 1 to 3 A method for integrated operation and control of photovoltaic and energy storage devices for microgrids, the method comprising the following steps:

[0075] Step S1: Obtain the current and voltage waveform data across the filter inductor of the photovoltaic energy storage device; perform frequency domain sweep on the current and voltage waveform data to generate complex impedance data of the filter inductor; based on the complex impedance data of the filter inductor, map the local saturation start point and nonlinear segment of the magnetic core to generate saturation characteristic data;

[0076] Step S2: Connect a bidirectional DC bias winding in parallel next to the filter inductor, and adjust the bias current based on the saturation characteristic data to generate core operating point adjustment data; split the filter into two groups, main and auxiliary, and dynamically switch them in parallel / series under different operating conditions using solid-state relays, and configure the filter impedance based on the saturation characteristic data to generate filter impedance configuration data.

[0077] Step S3: Adaptively adjust the PWM carrier slope and dead zone of the photovoltaic energy storage device using the magnetic core operating point adjustment data and filter impedance configuration data to obtain PWM modulation parameter optimization data; perform virtual inductance / resistance compensation superposition on the PWM modulation parameter optimization data to generate adaptive compensation control data;

[0078] Step S4: Generate control commands from the adaptive compensation control data, and input the generated control commands into the preset microgrid for command-coordinated control, so as to perform integrated operation control optimization of photovoltaic and energy storage equipment for microgrid.

[0079] This invention acquires the current and voltage waveform data across the filter inductor of a photovoltaic energy storage device and performs frequency domain sweep analysis to accurately obtain the complex impedance characteristics of the filter inductor. This allows for the mapping of the local saturation start point and nonlinear segment of the magnetic core, achieving precise modeling of the magnetic core saturation characteristics. Furthermore, a bidirectional DC bias winding is connected in parallel with the filter inductor, and the magnetic core saturation characteristics are dynamically adjusted, enabling controllable adjustment of the magnetic core operating point. By dividing the filter inductor into two groups, main and auxiliary, and using solid-state relays to achieve intelligent parallel / series switching under different operating conditions, the equivalent impedance of the filter can be flexibly configured under complex operating conditions, enhancing the electromagnetic compatibility and stability of the system. Based on this, by combining the magnetic core operating point adjustment data and the filter impedance configuration data, the PWM carrier slope and dead zone are adaptively adjusted. This not only optimizes the PWM modulation strategy but also effectively suppresses voltage and current distortion through the superposition of virtual inductance and virtual resistance compensation. The resulting adaptive compensation control command can work collaboratively with other control units in the microgrid, significantly improving the operational stability, response flexibility, and overall energy efficiency of the photovoltaic energy storage device in the microgrid. Therefore, this invention solves the problems of unstable filtering performance, fixed modulation parameters, and insufficient system coordination in traditional photovoltaic energy storage equipment by accurately identifying the nonlinear characteristics of the filter inductor, dynamically adjusting the operating point of the magnetic core, adaptively optimizing PWM modulation parameters, and coordinating with the microgrid.

[0080] In this embodiment of the invention, reference Figure 1 The diagram shown is a flowchart illustrating the steps of a comprehensive operation and control method for photovoltaic and energy storage devices in a microgrid according to the present invention. In this example, the comprehensive operation and control method for photovoltaic and energy storage devices in a microgrid includes the following steps:

[0081] Step S1: Obtain the current and voltage waveform data across the filter inductor of the photovoltaic energy storage device; perform frequency domain sweep on the current and voltage waveform data to generate complex impedance data of the filter inductor; based on the complex impedance data of the filter inductor, map the local saturation start point and nonlinear segment of the magnetic core to generate saturation characteristic data;

[0082] Step S2: Connect a bidirectional DC bias winding in parallel next to the filter inductor, and adjust the bias current based on the saturation characteristic data to generate core operating point adjustment data; split the filter into two groups, main and auxiliary, and dynamically switch them in parallel / series under different operating conditions using solid-state relays, and configure the filter impedance based on the saturation characteristic data to generate filter impedance configuration data.

[0083] Step S3: Adaptively adjust the PWM carrier slope and dead zone of the photovoltaic energy storage device using the magnetic core operating point adjustment data and filter impedance configuration data to obtain PWM modulation parameter optimization data; perform virtual inductance / resistance compensation superposition on the PWM modulation parameter optimization data to generate adaptive compensation control data;

[0084] Step S4: Generate control commands from the adaptive compensation control data, and input the generated control commands into the preset microgrid for command-coordinated control, so as to perform integrated operation control optimization of photovoltaic and energy storage equipment for microgrid.

[0085] In this embodiment of the invention, instantaneous voltage and current waveforms at both ends of the filter inductor are acquired using high-precision data acquisition devices (such as an isolation voltage probe and a shunt current detector). The data sampling frequency is no less than 500kHz to ensure that high-frequency harmonic components are not lost. The acquired data is sent to an embedded digital signal processing unit, where short-time Fourier transform (STFT) is used for frequency domain sweep processing to calculate the impedance amplitude and phase at different frequency points, generating complex impedance data for the filter inductor. By combining the BH curve parameter table provided by the core material or experimental measurement curves, the phase delay abrupt change region and amplification anomaly point in the complex impedance data are compared to determine the initial point of the core entering the saturation region, i.e., the local saturation start point. Combined with the nonlinear response segment, the range of the nonlinear segment is defined. The final output includes saturation characteristic data containing the start point, response curvature change, and nonlinear segment threshold. A bidirectional DC bias winding is configured in the parallel channel of the filter inductor, and a controllable DC bias current is injected into the bias winding using a current source control method. The DC current is adjusted in real time by the DSP control module based on the magnetic flux density control threshold before the magnetic core enters saturation from the saturation characteristic data. It is generally controlled between 60% and 85% of the rated magnetic flux density of the magnetic core, generating core operating point adjustment data for core bias control. Simultaneously, the filter topology is split from the traditional fixed structure into two paths: a main filter group and an auxiliary filter group, with solid-state relays (SSRs) connected in series / parallel between each path. The control unit dynamically switches the main and auxiliary filter group combination according to the operating conditions of the photovoltaic energy storage device (e.g., charge / discharge switching, current fluctuation frequency) to respond to impedance requirements under different operating modes. By analyzing the impedance change characteristics in the saturation characteristic data, adaptive impedance distribution configuration is performed to generate filter impedance configuration data. The aforementioned core operating point adjustment data and filter impedance configuration data are input to the PWM control module. The control module dynamically adjusts the PWM carrier slope (ranging from 100V / µs to 500V / µs) and dead time (ranging from 0.1µs to 1.5µs) based on the current load state and current waveform, forming an adaptive PWM parameter combination for different core responses and filter path impedance states, outputting optimized PWM modulation parameter data. Based on this, a virtual inductor / resistor compensation algorithm is introduced using digital signal simulation. The required additional inductance or resistance is calculated using the current output voltage and current sampling values, simulating the inductive reactance response and energy consumption characteristics of the device, adjusting the modulation waveform in real time to achieve dynamic compensation for the system response, outputting adaptive compensation control data. This adaptive compensation control data is input to the control strategy generation unit, which, combined with the microgrid operating status (such as photovoltaic-storage conversion mode, grid-connected / off-grid status, voltage load requirements, etc.), generates corresponding PWM control logic instructions and device operating mode switching instructions, constructing a complete control instruction set. This control instruction set is input to the microgrid centralized control platform via a communication interface (such as CAN bus or RS485).Based on its own topology and demand response strategy, the platform synchronously distributes commands to the photovoltaic (PV) side, energy storage side, and inverter unit in real time, enabling precise and coordinated control within the PV-storage system. Ultimately, it completes the comprehensive operation and control optimization of PV-storage equipment to meet the operational needs of the microgrid, making equipment responses more intelligent, coordinated, and efficient.

[0086] As an example of the present invention, reference is made to Figure 2 As shown, in this example, step S1 includes:

[0087] Step S11: Use a current probe and a differential voltage probe to synchronously acquire transient waveforms on the input and output sides of the filter inductor to obtain current and voltage waveform data across the filter inductor of the photovoltaic energy storage device.

[0088] Step S12: Perform a fast Fourier transform on the current and voltage waveform data to generate frequency domain swept frequency response spectrum data; calculate the amplitude-phase ratio based on the frequency domain swept frequency response spectrum data to generate the complex impedance data of the filter inductor;

[0089] Step S13: Perform inverse admittance phasor transformation analysis on the complex impedance data of the filter inductor, extract the equivalent permeability nonlinear inflection point, and generate the starting point data of local saturation of the magnetic core; based on the starting point data of local saturation of the magnetic core, perform segment trend slope segmentation on the complex impedance data of the filter inductor, identify the nonlinear gain interval, and generate the nonlinear segment data of the magnetic core.

[0090] Step S14: Merge the local saturation start point data of the magnetic core with the nonlinear section data of the magnetic core to generate saturation characteristic data.

[0091] In this embodiment of the invention, during the operation of the photovoltaic-storage system, a high-bandwidth current probe (such as Tektronix TCPA300+TCP312A, with a bandwidth of not less than 100MHz) and a differential voltage probe (such as P5205A, with a bandwidth of not less than 100MHz) are simultaneously connected to the input and output terminals of the filter inductor to acquire transient current and voltage waveforms. The probes need to be connected to a high-speed digital oscilloscope, with a sampling rate set to at least 1MS / s, and a sampling window length covering at least one complete PWM cycle (e.g., for 10 kHz PWM, a window of at least 100μs is set). The system timestamp synchronization mechanism ensures that the waveform data is aligned in the time domain, ultimately outputting current and voltage waveform data sequences. The acquired current and voltage time-domain data are sent to the main control platform or FPGA signal processing unit, and frequency domain conversion is performed using the Fast Fourier Transform (FFT) algorithm. The FFT operation should cover from the fundamental frequency to at least the 20th harmonic component (e.g., for a 10kHz fundamental frequency, the analysis bandwidth needs to be extended to at least 200kHz) to comprehensively extract the impedance behavior under the influence of harmonics. Based on the transformation results, voltage amplitude and phase data and current amplitude and phase data at each frequency point are extracted. Complex ratios are then performed on the complex voltage and current at the same frequency point to calculate the frequency domain complex impedance data of the filter inductor, in ohms, containing both real (equivalent resistance) and imaginary (equivalent reactance) parts. Inverse admittance phasor transformation analysis is then performed on the generated complex impedance data. Specifically, after taking the reciprocal of the impedance to obtain the admittance, the real (conductance) and imaginary (susceptance) parts of the admittance are expanded along the frequency axis, and the curve of the admittance phasor versus frequency is fitted to analyze its local slope changes. In the admittance phasor diagram, the first abrupt change point from a linear to a nonlinear state of the magnetic core is identified. This abrupt change point is defined as the starting frequency region where the magnetic core enters local saturation, corresponding to an increase in the rate of change of impedance characteristics or the appearance of an inflection point. The frequency and corresponding voltage and current amplitude data at this point are extracted to generate data on the starting point of local saturation of the magnetic core. Based on this, the complex impedance curve is segmented, and the nonlinear response segment is defined according to the range of a significant upward trend in the slope of the local first derivative. This response segment exhibits an abnormal increase in the real or imaginary part of the impedance as the frequency increases or the waveform changes, outputting data for the nonlinear segment of the magnetic core. The aforementioned data on the starting point of local core saturation is fused with the data for the nonlinear segment of the magnetic core to construct a core saturation characteristic structure. This structure includes: the voltage amplitude corresponding to the starting frequency (in volts), the current amplitude corresponding to the starting frequency (in amperes), the frequency range of the nonlinear segment (in Hz), and the slope variation range of the nonlinear segment (in Ω / Hz). This structure constitutes the saturation characteristic data, used as the parameter input basis for subsequent core operating point adjustment and filter configuration modules. The saturation characteristic data should be stored in a non-volatile memory (such as EEPROM) within the controller for easy retrieval and association with dynamic control logic.

[0092] Preferably, in step S13, performing inverse admittance phasor transformation analysis on the complex impedance data of the filter inductor to extract the equivalent permeability nonlinear inflection point includes:

[0093] The complex impedance data of the filter inductor is conjugate normalized to generate unit amplitude admittance spectrum data.

[0094] The imaginary admittance curve is extracted from the unit amplitude admittance spectrum data to generate the inductive admittance change trajectory;

[0095] The slope change rate at local extreme points of the inductive admittance change trajectory is analyzed to generate a first-order slope derivative curve.

[0096] Identify abrupt slope changes in the first-order slope derivative curve, extract the inflection frequency points of magnetic permeability drops, and generate data on abrupt changes in magnetic permeability.

[0097] By combining the rate of change of complex impedance phase angle corresponding to the abrupt change frequency, the phase lag characteristic is verified, the key frequency points that conform to the core saturation response are output, and the data of the local saturation start point of the core are generated.

[0098] In this embodiment of the invention, the complex impedance data of the filter inductor obtained from step S12 (denoted as Z(f) = R(f) + jX(f), where...) A conjugate normalization operation is performed on the frequency. This operation aims to eliminate the influence of energy amplitude differences across different frequency bands, allowing the analysis to focus on the relative admittance characteristic variation trend. The normalization steps are as follows: Calculate the admittance Y(f) = 1 / Z(f) = G(f) + jB(f), where G(f) is the conductance and B(f) is the susceptance (i.e., the inductive admittance component). For each frequency point, the admittance magnitude |Y(f)| is normalized to its maximum value, yielding the unit amplitude admittance spectrum: Ynorm(f) = Y(f) / maxf |Y(f)|. The generated Ynorm(f) = Gnorm(f) + jBnorm(f) is the unit amplitude admittance spectrum data. Next, the imaginary part Bnorm(f), i.e., the trajectory of inductive admittance variation with frequency, is extracted from the unit amplitude admittance spectrum data. This trajectory reflects the inductive reactance variation behavior of the magnetic core in response to high frequencies and is the fundamental data for identifying saturation trends. Then, the first derivative of the imaginary part trajectory of the admittance is calculated using the central difference method (or a higher-order numerical difference method), and its first-order slope derivative curve is constructed: The derivative curve is then smoothed using a Savitzky-Golay filter (a window size of 5-7 points is recommended) to suppress high-frequency noise interference. Local extrema (i.e., locations of abrupt changes in the derivative slope) are identified in the smoothed first derivative curve. These extrema typically reflect abrupt changes in the core material's response. Combined with the original frequency coordinates, the frequency points corresponding to these locations are extracted as data for abrupt changes in permeability. Subsequently, the original complex impedance phase angle corresponding to the abrupt change data is... The phase change rate dθ(f) / df is extracted and calculated. If the phase hysteresis rate at the abrupt change point increases significantly (e.g., more than 3 times the average value of the previous interval), it indicates that the magnetic flux response delay at this frequency point is aggravated, indicating that the magnetic core has entered the nonlinear saturation region. Finally, the frequency point that simultaneously meets the following two conditions is marked as the starting point of local saturation of the magnetic core: a sudden change in the local extremum of the inductive admittance slope derivative curve; and the phase hysteresis change rate is higher than the set hysteresis threshold (the empirical threshold is recommended to be >0.5 rad / kHz). The additional data such as the complex impedance amplitude, phase angle, and admittance corresponding to the frequency point that meets the above conditions are encapsulated into a structured output and marked as the starting point data of local saturation of the magnetic core. This data is used by the subsequent control system to determine the critical range of the magnetic core operating point and supports PWM parameter optimization and filter dynamic configuration tasks.

[0099] Preferably, step S2 includes the following steps:

[0100] Step S21: Connect a bidirectional DC bias winding next to the filter inductor, and set the initial bias value of the adjustable constant current source based on the local saturation start point data of the magnetic core to generate the initial excitation parameters of the bias winding.

[0101] Step S22: Analyze the bias current regulation response curve of the saturation characteristic data to generate the gain coefficient curve of the bias current to the change of magnetic permeability; based on the gain coefficient curve of the bias current to the change of magnetic permeability, dynamically set the bias winding control command and generate core operating point adjustment data.

[0102] Step S23: Split the filter inductor into a main filter branch and an auxiliary filter branch, and configure independent sampling channels to measure the voltage and current of the two branches respectively, generating filter branch operating status monitoring data;

[0103] Step S24: Obtain the operating load data of the main filter branch and the auxiliary filter branch; based on the operating load data, identify the load disturbance mode of the filter branch operating status monitoring data and generate dynamic switching threshold condition data.

[0104] Step S25: Configure solid-state relay control logic, perform series-parallel dynamic switching of main and auxiliary filter branches based on dynamic switching threshold condition data and solid-state relay control logic, and generate filter topology switching status data; perform filter impedance mapping analysis on topology switching status data and saturation characteristic data to generate filter impedance configuration data under the current topology.

[0105] In this embodiment of the invention, a set of bidirectional DC bias windings is integrated next to the filter inductor winding structure. The winding wire diameter is selected as 0.5mm, and the number of coil turns is set to 100. The magnetic coupling structure ensures that the magnetic flux of the windings interacts significantly with the main winding. One end of the bias winding is connected to an adjustable constant current source module. The initial bias current setting range is ±500mA, and the adjustment resolution is 10mA. The constant current source module receives the bias control parameters through a digital signal control channel. Using the local saturation starting point frequency of the magnetic core output in step S13, corresponding to its saturation permeability critical point, the required magnetic flux density value at that point is calculated, converted into the initial excitation current value, and the initial excitation parameters of the bias winding are generated, which are used as the constant current output setting value under the system initialization control conditions. The constant current source bias current is scanned step by step (each step increment is 50mA), and the complex impedance response of the filter inductor is collected under each bias current. The frequency response data is recorded by a frequency domain sweep instrument. The permeability curves under various bias currents are fitted to construct a "bias current – ​​equivalent permeability" relationship curve, i.e., a gain coefficient curve. The segment with a gain coefficient slope greater than 0.2 μH / mA is identified as the permeability adjustment sensitive area. During actual operation, based on this curve, the system updates the control bias current command value in real time according to the feedback permeability change trend. A dynamic control function is set and output to a constant current source to achieve dynamic adaptive adjustment of the core operating point. The final output bias control current command is the core operating point adjustment data. The original single-unit filter inductor is split into a main filter branch and an auxiliary filter branch. Each branch is connected to an independent current / voltage sampling module. The sampling module uses a 16-bit ADC channel with a sampling frequency of 100kHz. The instantaneous AC current and voltage waveforms of the two branches are recorded respectively. The equivalent impedance, phase angle, amplitude change, and other operating state characteristics are calculated by the processing unit to generate operating state monitoring data for the two branches, including multiple dimensions such as impedance modulus, apparent power, and harmonic content. The system control unit acquires the power data, output frequency range, and inverter switching frequency variation trend of the current operating load as operating load data. This data is then jointly analyzed with the operating status data of the main and auxiliary filter branches. A disturbance identification method based on decision tree logic is used to classify and train the data features (e.g., identifying sudden changes in output power, step-by-step load switching, harmonic abrupt changes, etc.). After identifying the corresponding operating characteristic changes, these key inflection points are extracted as dynamic switching threshold condition data, including: current slope greater than 5A / ms, voltage abrupt change exceeding 10V / ms, or apparent power fluctuation greater than 20%. Based on the above threshold condition data, a control logic table for solid-state relays is set. For example, when a load abrupt change event is triggered (meeting any of the above conditions), the system triggers a switch from parallel to series connection of the auxiliary filter branch. A bidirectional solid-state relay module is used, with each branch configured with an independent control line, resulting in a response time of less than 1ms. The relays are managed by an FPGA controller, executing state transitions through a preset state transition diagram. The relay state record generates filter topology switching state data.Simultaneously, the switching state data is input into the filter parameter mapping module, and together with the current saturation characteristic data of the magnetic core, the equivalent impedance of the filter under the current topology is calculated. The output is the filter impedance configuration data, which is then read and optimized by the subsequent PWM controller.

[0106] Preferably, step S24, which involves identifying load disturbance patterns in the filter branch operating status monitoring data based on the load data, includes:

[0107] Perform periodic stability assessment on the load data and generate load period drift characteristic data;

[0108] Extract the power fluctuation threshold from the filter branch operating status monitoring data to obtain power fluctuation mutation index data;

[0109] Joint pattern classification is performed on load cycle drift characteristic data and power fluctuation mutation index data to generate disturbance pattern classification label data, which includes high-frequency impact type, intermittent oscillation type and continuous jitter type.

[0110] Mapping rules are set for the disturbance pattern classification label data to generate disturbance response priority data;

[0111] Configure event trigger thresholds based on disturbance response priority data, and generate dynamic switching threshold condition data.

[0112] In this embodiment of the invention, load current data and output active power data of the photovoltaic storage device during actual operation are obtained through a periodic stability discrimination operation, and the analysis period window is set to 500 milliseconds. A Fast Fourier Transform (FFT) algorithm is used to perform spectral analysis on the power data, identifying the dominant frequency component and extracting periodic distribution characteristics. If the dominant frequency offset exceeds 0.8 Hz or the periodic fluctuation amplitude is greater than 10% within three consecutive periods, it is determined to be a periodic drift phenomenon. Based on the time window, the offset intensity, frequency fluctuation range, and change trend are recorded to form load periodic drift characteristic data. Next, the current amplitude, voltage amplitude, and instantaneous power value are extracted from the filter branch operating status monitoring data, and a moving average is performed to calculate the active power change amplitude at 50 ms intervals. A mutation threshold of 15% of the rated output power is set. If the power fluctuation exceeds the threshold within 100 ms, this point is recorded as the "mutation moment," and the mutation direction and intensity are marked to form power fluctuation mutation index data. If the mutation intensity exceeds 25% of the rated power, a "high-intensity mutation" mark is added. Using the aforementioned load cycle drift characteristic data and power fluctuation mutation index data as input features, a joint classification model based on Support Vector Machine (SVM) is trained to identify the disturbance pattern types. The classification output labels include: high-frequency impact type (e.g., high-frequency sudden load caused by frequent mode switching in a photovoltaic-storage system); intermittent oscillation type (e.g., frequent switching of filtering tasks due to intermittent load fluctuations); and continuous jitter type (e.g., continuous low-amplitude disturbances caused by long-term unstable load conditions). The model output generates corresponding disturbance pattern classification label data. For different disturbance patterns, the system predefines a disturbance response priority mapping rule: high-frequency impact type disturbances have the highest priority, followed by continuous jitter type, and finally intermittent oscillation type. According to this mapping logic, each label is assigned a numerical priority identifier (e.g., 1 for highest, 3 for lowest), and the output is disturbance response priority data. Finally, based on the disturbance response priority data, corresponding event triggering threshold conditions are set. For example, when a high-frequency impulsive disturbance occurs, the threshold is set to a current surge rate exceeding 6A / ms; for a continuous jitter disturbance, the threshold is set to a voltage fluctuation amplitude greater than 12V lasting for more than 1 second; and for an intermittent oscillation disturbance, the threshold is set to a periodic variation exceeding 5% and a power surge amplitude between 10% and 15%. Based on these settings, the final dynamic switching threshold condition data is output and used as a reference for triggering dynamic topology switching in the solid-state relay control logic. Through this multi-level identification and classification mechanism, it can be ensured that the filtering system can respond quickly, switch appropriately, and operate stably when facing different disturbance scenarios.

[0113] Preferably, joint pattern classification of load cycle drift characteristic data and power fluctuation abrupt change index data includes:

[0114] Sliding cross-spectral focusing analysis is performed on load cycle drift characteristic data to generate frequency perturbation coupling index data;

[0115] The perturbation edge acceleration features of the power fluctuation abrupt change index data are extracted to obtain the abrupt change response intensity data;

[0116] A two-factor perturbation feature matrix is ​​constructed based on frequency perturbation coupling index data and mutation response intensity data. Non-convex fuzzy clustering is then performed to generate a perturbation distribution label map.

[0117] The regional density gradient projection of the disturbance distribution label map is calculated, and the projection morphology is determined to obtain disturbance pattern classification label data. Specifically, the projection morphology is determined as follows: when the projection morphology is a cluster-oblique impact type, it is marked as a high-frequency impact type; when the projection morphology is a divergence-loop type, it is marked as an intermittent oscillation type; and when the projection morphology is an expansion-resonance type, it is marked as a continuous shaking type.

[0118] In this embodiment of the invention, frequency drift trajectories within each cycle of load cycle drift characteristic data are extracted and converted into a time-frequency two-dimensional matrix. Using a time window of 250ms and a step size of 50ms, sliding cross-spectral focusing processing is performed on this two-dimensional frequency sequence. This processing, based on wavelet packet decomposition combined with cross-power spectral density function, quantifies the coupling degree of multiple frequency components in the same load cycle data, generating frequency perturbation coupling index data. This index reflects the abrupt change in frequency coupling per unit time, with a value ranging from 0.1 to 2.5; a higher value indicates a stronger perturbation focusing effect. Next, the fluctuation edges in the power fluctuation abrupt change index data are extracted, and cubic spline interpolation is used to fit the power waveform within 50ms before and after the abrupt change point. The second derivative of the fitted curve at the abrupt change edge point is calculated to obtain the perturbation edge acceleration characteristics. This feature is used to measure the intensity of the abrupt change response, with the numerical unit being watts per millisecond squared (W / ms²). The extracted features are normalized to constitute abrupt change response intensity data. Subsequently, a two-dimensional two-factor perturbation feature matrix is ​​constructed, where each perturbation trajectory corresponds to a feature point composed of the frequency perturbation coupling index and the abrupt response intensity. A non-convex fuzzy clustering algorithm based on a membership decay model is then applied to the feature matrix, with the number of clusters set to 3 and the iterative convergence threshold set to 1×10⁻⁶. -5After clustering, a perturbation distribution label map is output, where each perturbation sample is labeled as belonging to the same perturbation type based on the clustering results. The generated perturbation distribution label map is then further subjected to regional density gradient projection analysis. Specifically, kernel density estimation is performed along the diagonal direction in the two-dimensional perturbation space to construct isodense gradient lines; then, one-dimensional projection is performed on the isodense lines, projecting the density gradient onto the principal component axis direction to form a regional density gradient projection image. Subsequently, morphological discrimination is performed on this projection image, with the following logic: when a sharp peak is observed in the density center area, with a rapid downward trend on both sides, and the overall shape is a single-core diagonally extended slope, it is judged as a clustered-oblique-impact type, corresponding to a high-frequency impact type perturbation; when the projection shape shows multiple discontinuous peak-trough structures, exhibiting periodic divergence and return-convergence characteristics, it is judged as a divergence-loop type, corresponding to an intermittent oscillation type perturbation; when the projection shape shows a slow stretching, blurred boundaries, and increasing edge density oscillation trend, it is judged as an expansion-resonance type perturbation, corresponding to a continuous jitter type perturbation. Based on the above discrimination results, the perturbation mode classification label data is output, in which the perturbation mode type of each sample is marked, and the classification label is used as the key reference information for subsequent dynamic topology switching logic and control priority setting.

[0119] Preferably, step S3, which involves adaptively adjusting the PWM carrier slope and dead zone of the photovoltaic energy storage device using magnetic core operating point adjustment data and filter impedance configuration data, includes:

[0120] Identify the nonlinear saturation range of the magnetic core operating point adjustment data;

[0121] Perform frequency response matching analysis on the filter impedance configuration data to generate dynamic impedance characteristic data of the filter.

[0122] Based on the nonlinear saturation range, the dynamic impedance characteristic data of the filter are used to generate PWM carrier slope adaptive adjustment rules, and the carrier slope adjustment strategy data is obtained.

[0123] Based on carrier slope adjustment strategy data, dead zone correction is performed on the optical storage device to generate dead zone time correction parameter data.

[0124] The carrier slope adjustment strategy data and dead time correction parameter data are jointly mapped to generate PWM modulation parameter optimization data.

[0125] In this embodiment of the invention, the excitation current-magnetic flux density curve data from the core operating point adjustment data is retrieved, and the first derivative is calculated using the five-point sliding differential method to obtain the permeability variation trend. The saturation threshold is set as a permeability decrease rate greater than 5 × 10⁻⁶. - ³H - ¹A -The interval ¹ is defined as a nonlinear saturation interval when three or more consecutive data points within this interval meet the threshold condition. The output data format is the start and end time of the interval (unit: ms) and the corresponding current amplitude range (unit: A), constituting the nonlinear saturation response interval data. The filter inductor and capacitor configuration parameters under the current topology are obtained, and combined with the actual PWM modulation frequency, frequency domain analysis is performed on the filter impedance, covering a frequency range of 1kHz to 100kHz, with a step of 100Hz. The equivalent impedance is calculated, and the impedance magnitude and phase angle are recorded for each frequency point to form a frequency response data table. The envelope features of the response curve are extracted through Hilbert transform, and overlap analysis is performed with the operating frequency of the nonlinear saturation interval to obtain the dynamic impedance characteristic data of the filter in the key frequency range. Based on the intersection region of the saturation interval and the impedance frequency response, the carrier frequency is divided into a low response region (<5kHz), a medium response region (5–20kHz), and a high response region (>20kHz), with different carrier slope ranges set for each region. The slope adjustment rules were set through experimental calibration as follows: the slope in the low response region was set to 0.2V / μs; the slope in the medium response region was set to 0.5V / μs; and the slope in the high-frequency response region was set to 1.0V / μs. Based on the segment where the magnetic core saturation start frequency point is located, the corresponding slope level was matched, and the carrier frequency modulation step was limited to ±0.5kHz to generate carrier slope adjustment strategy data for subsequent PWM control module scheduling. Based on the slope adjustment strategy, the switching delay characteristics of the main power switching devices were obtained (e.g., MOSFET rise / fall delay is generally set to 20ns~50ns), and the dead-time compensation requirement was calculated in conjunction with the current rise slope. The dead time Dt was set as follows: Dt=max{Ton_delay, Toff_delay}+Δt_margin, where Δt_margin is fixed at a safety margin of 10ns. The final dead-time correction parameter range was controlled between 150ns and 400ns to ensure that conduction overlap was avoided. A parameter optimization mapping table is established, employing a binary table matching method to create a combined index between slope levels (e.g., low / medium / high) and corresponding dead times. The output PWM modulation parameters include: modulation frequency (kHz), rise / fall slope (V / μs), dead time (ns), and corresponding enable status word. All parameters are encapsulated as structured configuration data and dynamically loaded into the PWM drive module via a digital signal processor (DSP) to form the final optimized PWM modulation parameters. This enables highly adaptive PWM modulation control of the optical storage device under dynamic core conditions and filtering conditions.

[0126] Preferably, dead-zone correction of the photovoltaic storage device based on carrier slope adjustment strategy data includes:

[0127] Model the switching delay caused by dead zone on the carrier slope adjustment strategy data to generate switching delay error modeling data;

[0128] The dead time correction amount is calculated based on the switching delay error model data, and the dead time correction amount data is generated.

[0129] The dead time width of the optical storage device is corrected using the dead time correction data, and the corrected dead time width is adjusted by feedback to generate dead time correction parameter data.

[0130] In the embodiments of the present invention, by adjusting the carrier slope strategy data that has been generated (where the slope is generally controlled within 0.2 - 1.0 V / μs), discrete point modeling is performed on the turn-on delay (Ton_delay) and turn-off delay (Toff_delay) of the main power switch devices (such as SiC MOSFET or IGBT) used in the energy storage device under different slope conditions. During the test, a high-speed sampling oscilloscope (with a bandwidth not less than 100 MHz and a sampling rate greater than 500 MSa / s) is used to collect the driving waveforms, and the actual time difference between the slope input signal and the device response is recorded. At least 50 measurements are made under each slope condition, and their mean and standard deviation are calculated. Taking the slope as the horizontal axis and the delay as the vertical axis, the turn-on and turn-off delay curves are respectively fitted using a second-order polynomial fitting function: Ton_delay(s) = a1·S² + b1·S + c1; Toff_delay(s) = a2·S² + b2·S + c2; where S is the carrier slope (unit: V / μs), a1, b1, c1 are the turn-on delay coefficients, and a2, b2, c2 are the turn-off delay coefficients. Finally, the switching delay error modeling data is formed and stored in the form of a table or function for subsequent dead-time correction amount calculation. According to the current carrier slope value S0 (such as 0.6 V / μs) selected in the control system, the switching delay error model is called to calculate the corresponding Ton_delay(S0) and Toff_delay(S0) respectively. In the system default configuration, the initial dead-time Dt0 is set to 300 ns. Compare Ton_delay and Toff_delay in the simulation or actual measurement with Dt0. If Dt0 < max(Ton_delay, Toff_delay), it is determined that there is a risk of overlapping conduction, and the dead-time correction amount ΔDt needs to be calculated. The calculation formula is as follows: ΔDt = max{Ton_delay, Toff_delay} + Δt_margin - Dt0; where Δt_margin is the fixed safety margin time, with a value of 10 ns. If ΔDt is positive, it means that the dead-time needs to be extended; if it is negative, the dead-time can be shortened under the premise of ensuring safety to reduce power loss. Finally, the dead-time correction amount data is output, with the unit of ns, and the correction accuracy is controlled within ±5 ns. According to the minimum dead-time step granularity supported by the current PWM controller (such as 10 ns), the dead-time is set to Dt_adj = Dt0 + ΔDt, and this value is written into the PWM dead-time configuration register inside the DSP or FPGA. During actual operation, the following two types of feedback signals are monitored: The output voltage of the bridge arm is sampled using an ADC with a sampling accuracy not less than 200 ns to detect the overlapping area or blank window area between the two switchings and evaluate the error amplitude. The waveform of the bus current at the switching moment is recorded, and the peak intensity is extracted through a Savitzky-Golay filter to identify whether a short circuit or overshoot is caused by insufficient dead-time setting.If the above feedback indicators deviate from the set threshold (e.g., voltage mismatch exceeding ±10ns, or current spike exceeding the rated value by 20%), the dead time is adjusted by ±10ns according to the feedback direction and rewritten to the PWM control register. After finally converging to the optimal dead time configuration, dead time correction parameter data is output, including the final set value (e.g., 350ns), adjustment history, corresponding carrier slope parameters, etc., for subsequent strategy fusion and adaptive control optimization.

[0131] As an example of the present invention, reference is made to Figure 3 As shown, step S4 in this example includes:

[0132] Step S41: Perform dynamic instruction encoding on the adaptive compensation control data to generate multi-dimensional control instruction set data;

[0133] Step S42: Identify instruction redundancy conflicts in the multidimensional control instruction set data, eliminate redundant and mutually exclusive instructions, and thus generate deduplication and optimization data for control instructions;

[0134] Step S43: Input the deduplication and optimization data of the control commands into the microgrid power routing module to perform regional load response mapping and generate microgrid command matching mapping data;

[0135] Step S44: Calculate the control timing consistency of the microgrid command matching mapping data to obtain the collaborative control scheduling sequence data; input the collaborative control scheduling sequence data to the preset microgrid control interface for command distribution and execution, generate operation feedback response data, and perform integrated operation control optimization operation for microgrid-oriented photovoltaic and energy storage equipment.

[0136] In this embodiment of the invention, an embedded encoding and decoding framework is used to call the instruction encoding module within the controller's local DSP or FPGA system to convert adaptive compensation control data (including PWM slope parameters, dead-time correction parameters, inductor virtual compensation values, and dynamic resistance adjustment amounts) into standardized control fields. The specific field definitions are as follows: CMD_TYPE (1 byte): Control command type, such as 01 representing modulation control and 02 representing current control; PARAM_ID (1 byte): Parameter identifier, such as A1 representing carrier slope; VALUE (4 bytes): Actual parameter value, encoded using IEEE 754 floating-point numbers; PRIORITY (1 byte): Priority flag, value range 0-15; TIME_TAG (2 bytes): Relative timing identifier for instruction issuance, unit ms. The above structures are combined into instruction units, organized into instruction sets according to the device structure topology (inverter module, battery module, filter module), forming multi-dimensional control instruction set data, where dimensions include module dimension, function dimension, timing dimension, and parameter dimension. All instruction data is checked using CRC16 and cached in a high-priority interrupt queue. For the multi-dimensional control instruction set output in step S41, a preset conflict identification rule base is loaded, and the following two types of checks are performed: traversing instructions with the same CMD_TYPE and PARAM_ID but different TIME_TAG, keeping the latest one and eliminating the rest; for parameter pairs that cannot be effective at the same time (such as carrier rise slope adjustment and carrier switching frequency dynamic jitter), the mutual exclusion table is consulted, and the one with higher priority or more urgent response time requirement is kept. During the processing, the instruction hash index mechanism is called to achieve O(1) level fast matching and comparison, and finally outputs the deduplicated and optimized control instruction data after sorting. Each record is accompanied by a redundancy processing mark and conflict handling description for record traceability. Taking the regional load response as the core, the mapping logic between control commands and microgrid nodes is realized through the microgrid power routing module. First, the node load distribution map and equipment scheduling capacity table of the current region are imported, and the parameters involved in the control command (such as the change value of filter compensation resistor or the battery charging and discharging limit value) are matched with the actual power distribution between nodes. Using an L×M spatial distribution matrix (where L is the number of nodes and M is the number of control types), the following matching operation is performed on each command: The power module mapping matrix is ​​queried based on CMD_TYPE; the ability to receive control commands is determined based on the current load factor of the equipment (e.g., converter current margin > 10%); if mapping is successful, a matching record is generated, including the target node ID, scheduling weight, delay compensation item, etc.; finally, the microgrid command matching mapping data is output, and the control path, power adjustment direction, and node reception status are marked in matrix form. Timing consistency analysis is performed on the execution order of multiple control commands.Based on the response delay (τ) of each node and the expected execution time of the instruction (T_req), the following timing scheduling strategy is adopted: If |T_req–τ|> the maximum allowable deviation ΔT (e.g., 50ms), the scheduling time is corrected; a directed scheduling graph is constructed according to the dependency graph, and the instructions are topologically sorted; the sorting results are written into the scheduling sequence table to form coordinated control scheduling sequence data; then, this sequence data is sent to each control terminal through the microgrid communication interface (e.g., Modbus TCP / IP or IEC 61850 MMS) to actually issue instructions and start the equipment response. After all terminal devices complete the instruction execution, they return a status frame, including the execution success flag, execution time, feedback parameters, and status anomaly identification code, which are integrated into operation feedback response data for the dispatch center to review the control results and adjust the next round of scheduling.

[0137] This specification provides a comprehensive operation and control system for photovoltaic and energy storage devices in microgrids, used to execute the aforementioned comprehensive operation and control method for photovoltaic and energy storage devices in microgrids. This comprehensive operation and control system for photovoltaic and energy storage devices in microgrids includes:

[0138] The data acquisition module is used to acquire the current and voltage waveform data across the filter inductor of the photovoltaic energy storage device; to perform frequency domain sweep on the current and voltage waveform data to generate complex impedance data of the filter inductor; and to map the local saturation start point and nonlinear segment of the magnetic core based on the complex impedance data of the filter inductor to generate saturation characteristic data.

[0139] The data intervention configuration module is used to connect a bidirectional DC bias winding in parallel next to the filter inductor, and to adjust the bias current based on the saturation characteristic data to generate core operating point adjustment data; it also splits the filter into two groups, main and auxiliary, which are dynamically switched in parallel / series by solid-state relays under different operating conditions, and configures the filter impedance based on the saturation characteristic data to generate filter impedance configuration data.

[0140] The control optimization module is used to adaptively adjust the PWM carrier slope and dead zone of the photovoltaic energy storage device by using magnetic core operating point adjustment data and filter impedance configuration data to obtain PWM modulation parameter optimization data; and to perform virtual inductance / resistance compensation superposition on the PWM modulation parameter optimization data to generate adaptive compensation control data.

[0141] The instruction distribution module is used to generate control instructions from adaptive compensation control data and input the generated control instructions into a preset microgrid for instruction-coordinated control, so as to perform integrated operation control optimization of photovoltaic and energy storage equipment for microgrid.

[0142] The beneficial effects of this invention lie in the fact that by setting up a data acquisition module, high-precision acquisition of the voltage and current waveforms across the filter inductor of the photovoltaic energy storage device is achieved. Furthermore, by using frequency domain sweep technology to extract the complex impedance characteristics of the filter inductor, the local saturation start point and nonlinear segment of the magnetic core are identified, effectively constructing a high-resolution saturation characteristic model reflecting the dynamic saturation behavior of the magnetic core, providing precise physical support for subsequent control strategies. The data intervention configuration module, by introducing an adjustable DC bias winding and a magnetic core operating point adjustment mechanism, enhances the controllability and linear range of the magnetic core. It also supports dynamic reconfiguration and series-parallel switching of the main and auxiliary branches of the filter inductor under multiple operating conditions. Combined with saturation characteristic data, intelligent configuration of the filter impedance is performed, significantly improving the filtering performance. The invention enhances the equivalent impedance adaptation capability and response sensitivity of the device under complex operating conditions. A control optimization module integrates core operating point adjustment data and filter impedance configuration data to achieve dynamic adaptive adjustment of the PWM carrier slope and dead zone. Furthermore, virtual inductance / resistance compensation is superimposed, enabling fine-grained tuning of control parameters. This effectively reduces switching losses, harmonic distortion, and system response lag, improving the real-time performance and stability of the modulation strategy. The instruction distribution module supports intelligent encoding, conflict avoidance, and microgrid scenario matching of adaptive compensation control data, thereby achieving coordinated distribution and closed-loop feedback of control instructions. This significantly improves the control coordination, operating efficiency, and intelligent response level of the photovoltaic-storage equipment in the microgrid. Therefore, this invention solves the problems of unstable filtering performance, fixed modulation parameters, and insufficient system coordination in traditional photovoltaic-storage equipment by accurately identifying the nonlinear characteristics of the filter inductor, dynamically adjusting the core operating point, adaptively optimizing PWM modulation parameters, and implementing microgrid coordinated control.

[0143] Therefore, the embodiments should be considered as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of the equivalents of the application are intended to be included within the invention.

[0144] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A method for integrated operation and control of photovoltaic and energy storage devices for microgrids, characterized in that, Includes the following steps: Step S1: Obtain the current and voltage waveform data across the filter inductor of the photovoltaic energy storage device; perform frequency domain sweep on the current and voltage waveform data to generate complex impedance data of the filter inductor; based on the complex impedance data of the filter inductor, map the local saturation start point and nonlinear segment of the magnetic core to generate saturation characteristic data; Step S2: Connect a bidirectional DC bias winding in parallel next to the filter inductor, and adjust the bias current based on the saturation characteristic data to generate core operating point adjustment data; split the filter into two groups, main and auxiliary, and dynamically switch them in parallel / series under different operating conditions using solid-state relays, and configure the filter impedance based on the saturation characteristic data to generate filter impedance configuration data. Step S3: Adaptively adjust the PWM carrier slope and dead zone of the photovoltaic energy storage device using the magnetic core operating point adjustment data and filter impedance configuration data to obtain optimized PWM modulation parameter data; Virtual inductance / resistance compensation is superimposed on the PWM modulation parameter optimization data to generate adaptive compensation control data; wherein, step S3, which uses magnetic core operating point adjustment data and filter impedance configuration data to adaptively adjust the PWM carrier slope and dead zone of the photovoltaic energy storage device, includes: Identify the nonlinear saturation range of the magnetic core operating point adjustment data; Perform frequency response matching analysis on the filter impedance configuration data to generate dynamic impedance characteristic data of the filter. Based on the nonlinear saturation range, the dynamic impedance characteristic data of the filter is used to generate the PWM carrier slope adaptive adjustment rule, and the carrier slope adjustment strategy data is obtained. This includes dividing the carrier frequency into low response region, medium response region and high response region according to the intersection region of saturation range and impedance frequency response, and setting different carrier slope ranges for each region. Based on carrier slope adjustment strategy data, dead zone correction is performed on the optical storage device to generate dead zone time correction parameter data. Joint mapping is performed on carrier slope adjustment strategy data and dead time correction parameter data to generate PWM modulation parameter optimization data; among which, dead time correction of the photovoltaic storage device based on carrier slope adjustment strategy data includes: Model the switching delay caused by dead zone on the carrier slope adjustment strategy data to generate switching delay error modeling data; The dead time correction amount is calculated based on the switching delay error model data, and the dead time correction amount data is generated. The dead time width of the optical storage device is corrected using the dead time correction data, and the corrected dead time width is adjusted by feedback to generate dead time correction parameter data. Step S4: Generate control commands from the adaptive compensation control data, and input the generated control commands into the preset microgrid for command-coordinated control, so as to perform integrated operation control optimization of photovoltaic and energy storage equipment for microgrid.

2. The integrated operation and control method for photovoltaic and energy storage devices for microgrids according to claim 1, characterized in that, Step S1 includes the following steps: Step S11: Use a current probe and a differential voltage probe to synchronously acquire transient waveforms on the input and output sides of the filter inductor to obtain current and voltage waveform data across the filter inductor of the photovoltaic energy storage device. Step S12: Perform a fast Fourier transform on the current and voltage waveform data to generate frequency domain swept frequency response spectrum data; calculate the amplitude-phase ratio based on the frequency domain swept frequency response spectrum data to generate the complex impedance data of the filter inductor; Step S13: Perform inverse admittance phasor transformation analysis on the complex impedance data of the filter inductor, extract the equivalent permeability nonlinear inflection point, and generate the starting point data of local saturation of the magnetic core; based on the starting point data of local saturation of the magnetic core, perform segment trend slope segmentation on the complex impedance data of the filter inductor, identify the nonlinear gain interval, and generate the nonlinear segment data of the magnetic core. Step S14: Merge the local saturation start point data of the magnetic core with the nonlinear section data of the magnetic core to generate saturation characteristic data.

3. The integrated operation and control method for photovoltaic and energy storage devices for microgrids according to claim 2, characterized in that, In step S13, the admittance phasor inverse transformation analysis is performed on the complex impedance data of the filter inductor to extract the equivalent permeability nonlinear inflection points, including: The complex impedance data of the filter inductor is conjugate normalized to generate unit amplitude admittance spectrum data. The imaginary admittance curve is extracted from the unit amplitude admittance spectrum data to generate the inductive admittance change trajectory; The slope change rate at local extreme points of the inductive admittance change trajectory is analyzed to generate a first-order slope derivative curve. Identify abrupt slope changes in the first-order slope derivative curve, extract the inflection frequency points of magnetic permeability drops, and generate data on abrupt changes in magnetic permeability. By combining the rate of change of complex impedance phase angle corresponding to the abrupt change frequency, the phase lag characteristic is verified, the key frequency points that conform to the core saturation response are output, and the data of the local saturation start point of the core are generated.

4. The integrated operation and control method for photovoltaic and energy storage devices for microgrids according to claim 1, characterized in that, Step S2 includes the following steps: Step S21: Connect a bidirectional DC bias winding next to the filter inductor, and set the initial bias value of the adjustable constant current source based on the local saturation start point data of the magnetic core to generate the initial excitation parameters of the bias winding. Step S22: Analyze the bias current regulation response curve of the saturation characteristic data to generate the gain coefficient curve of the bias current to the change of magnetic permeability; based on the gain coefficient curve of the bias current to the change of magnetic permeability, dynamically set the bias winding control command and generate core operating point adjustment data. Step S23: Split the filter inductor into a main filter branch and an auxiliary filter branch, and configure independent sampling channels to measure the voltage and current of the two branches respectively, generating filter branch operating status monitoring data; Step S24: Obtain the operating load data of the main filter branch and the auxiliary filter branch; based on the operating load data, identify the load disturbance mode of the filter branch operating status monitoring data and generate dynamic switching threshold condition data. Step S25: Configure solid-state relay control logic, perform series-parallel dynamic switching of main and auxiliary filter branches based on dynamic switching threshold condition data and solid-state relay control logic, and generate filter topology switching status data; perform filter impedance mapping analysis on topology switching status data and saturation characteristic data to generate filter impedance configuration data under the current topology.

5. The integrated operation and control method for photovoltaic and energy storage devices for microgrids according to claim 4, characterized in that, Step S24, which involves identifying load disturbance patterns in the filter branch operating status monitoring data based on the load data, includes: Perform periodic stability assessment on the load data and generate load period drift characteristic data; Extract the power fluctuation threshold from the filter branch operating status monitoring data to obtain power fluctuation mutation index data; Joint pattern classification is performed on load cycle drift characteristic data and power fluctuation mutation index data to generate disturbance pattern classification label data, which includes high-frequency impact type, intermittent oscillation type and continuous jitter type. Mapping rules are set for the disturbance pattern classification label data to generate disturbance response priority data; Configure event trigger thresholds based on disturbance response priority data, and generate dynamic switching threshold condition data.

6. The integrated operation and control method for photovoltaic and energy storage devices for microgrids according to claim 5, characterized in that, Joint pattern classification of load cycle drift characteristic data and power fluctuation abrupt change index data includes: Sliding cross-spectral focusing analysis is performed on load cycle drift characteristic data to generate frequency perturbation coupling index data; The perturbation edge acceleration features of the power fluctuation abrupt change index data are extracted to obtain the abrupt change response intensity data; A two-factor perturbation feature matrix is ​​constructed based on frequency perturbation coupling index data and mutation response intensity data. Non-convex fuzzy clustering is then performed to generate a perturbation distribution label map. The regional density gradient projection of the disturbance distribution label map is calculated, and the projection morphology is determined to obtain disturbance pattern classification label data. Specifically, the projection morphology is determined as follows: when the projection morphology is a cluster-oblique impact type, it is marked as a high-frequency impact type; when the projection morphology is a divergence-loop type, it is marked as an intermittent oscillation type; and when the projection morphology is an expansion-resonance type, it is marked as a continuous shaking type.

7. The integrated operation and control method for photovoltaic and energy storage devices for microgrids according to claim 1, characterized in that, Step S4 includes the following steps: Step S41: Perform dynamic instruction encoding on the adaptive compensation control data to generate multi-dimensional control instruction set data; Step S42: Identify instruction redundancy conflicts in the multidimensional control instruction set data, eliminate redundant and mutually exclusive instructions, and thus generate deduplication and optimization data for control instructions; Step S43: Input the deduplication and optimization data of the control commands into the microgrid power routing module to perform regional load response mapping and generate microgrid command matching mapping data; Step S44: Calculate the control timing consistency of the microgrid command matching mapping data to obtain the collaborative control scheduling sequence data; input the collaborative control scheduling sequence data to the preset microgrid control interface for command distribution and execution, generate operation feedback response data, and perform integrated operation control optimization operation for microgrid-oriented photovoltaic and energy storage equipment.

8. A comprehensive operation and control system for photovoltaic and energy storage devices in microgrids, characterized in that, For executing the integrated operation control method for photovoltaic and energy storage devices for microgrids as described in claim 1, the integrated operation control system for photovoltaic and energy storage devices for microgrids includes: The data acquisition module is used to acquire the current and voltage waveform data across the filter inductor of the photovoltaic energy storage device; to perform frequency domain sweep on the current and voltage waveform data to generate complex impedance data of the filter inductor; and to map the local saturation start point and nonlinear segment of the magnetic core based on the complex impedance data of the filter inductor to generate saturation characteristic data. The data intervention configuration module is used to connect a bidirectional DC bias winding in parallel next to the filter inductor, and to adjust the bias current based on the saturation characteristic data to generate core operating point adjustment data; it also splits the filter into two groups, main and auxiliary, which are dynamically switched in parallel / series by solid-state relays under different operating conditions, and configures the filter impedance based on the saturation characteristic data to generate filter impedance configuration data. The control optimization module is used to adaptively adjust the PWM carrier slope and dead zone of the photovoltaic energy storage device by using magnetic core operating point adjustment data and filter impedance configuration data to obtain PWM modulation parameter optimization data; and to perform virtual inductance / resistance compensation superposition on the PWM modulation parameter optimization data to generate adaptive compensation control data. The instruction distribution module is used to generate control instructions from adaptive compensation control data and input the generated control instructions into a preset microgrid for instruction-coordinated control, so as to perform integrated operation control optimization of photovoltaic and energy storage equipment for microgrid.

Images