Multi-energy collaborative intelligent micro-grid control method, device, equipment and medium

Abstract

The application relates to a multi-energy collaborative intelligent micro-grid control method, device, equipment and medium. The method accurately evaluates the inertia demand of the system by collecting the micro-grid frequency signal in real time and calculating the change rate, dynamically calculates the maximum virtual inertia that can be provided by each unit in combination with the physical characteristics and operation constraints of each energy unit, dynamically matches the system demand and unit capacity to generate an optimal allocation scheme by using gray correlation analysis, drives each unit special control model to generate differentiated power regulation instructions, and finally executes power support through power electronic converters, thereby realizing accurate matching and efficient collaboration of multi-energy virtual inertia capacity, significantly enhancing the frequency stability of high-proportion new energy power grids, effectively avoiding resource allocation mismatch and poor control adaptability in traditional methods, and ensuring the dynamic safety of the system under complex operating conditions.

Description

Technical Field

[0001] This invention belongs to the field of power system stability control technology, and in particular relates to a multi-energy coordinated intelligent microgrid control method, device, equipment and medium. Background Technology

[0002] With the widespread application of renewable energy and power electronic equipment in microgrids, the physical inertia of microgrid systems has significantly decreased, leading to faster system response to disturbances and reduced stability margin. Traditional microgrids primarily rely on the rotational inertia provided by synchronous generators to maintain system frequency stability. However, the extensive use of power electronic converters in modern microgrids results in systems exhibiting "low inertia" and "weak damping" characteristics. Under these circumstances, when power disturbances occur, the frequency fluctuation amplitude increases, system stability decreases, and may even lead to faults such as protection activation, load shedding, and power source disconnection.

[0003] Traditional control methods based on the inertial characteristics of synchronous machines are ill-suited to this low inertia profile. Virtual inertial control, as an economical and effective solution, introduces virtual inertia and damping elements into the control system of power electronic devices, enabling them to exhibit inertial characteristics similar to those of synchronous generators. However, existing virtual inertial control methods are mostly designed for single-energy systems, lacking consideration for the coordinated use of multiple energy sources (such as wind power, photovoltaics, and energy storage) in microgrids, making it difficult to achieve balanced distribution and dynamic adjustment of system inertia. Summary of the Invention

[0004] Based on this, it is necessary to provide a multi-energy coordinated smart microgrid control method, device, equipment and medium to address the above-mentioned technical problems. The aim is to solve the low inertia problem in microgrids with a high proportion of power electronics. By coordinating virtual inertia control strategies of different types of power sources, the system inertia can be balanced and dynamically adjusted, thereby improving the system's transient stability and anti-disturbance capability.

[0005] Firstly, this application provides a multi-energy coordinated smart microgrid control method, including:

[0006] S1. The frequency signal of the microgrid system is collected by the synchronous phasor measurement unit deployed at the key nodes of the microgrid; based on the frequency change rate of the frequency signal, the required virtual inertia of the microgrid system is calculated through inertial demand assessment.

[0007] S2. Based on the physical model and operating constraints of each energy unit in the microgrid system, calculate the maximum virtual inertia that each energy unit can provide;

[0008] S3. Based on the demand virtual inertia and the maximum available virtual inertia, grey relational analysis is used to calculate the actual allocated virtual inertia of each energy unit.

[0009] S4. Based on the actual allocation of virtual inertia, generate additional power commands for each energy unit through the corresponding operation control model of each energy unit.

[0010] S5. The additional power commands corresponding to each energy unit are superimposed on the local control loop of the microgrid system to drive the power electronic converter to perform power regulation, so as to carry out virtual inertial control of multi-energy coordination.

[0011] Secondly, this application also provides a multi-energy coordinated smart microgrid control device for implementing the method described in the first aspect, the device comprising:

[0012] The dynamic inertial demand assessment module is used to collect the frequency signal of the microgrid system through synchronous phasor measurement units deployed at key nodes of the microgrid; based on the frequency change rate of the frequency signal, the virtual inertia demand of the microgrid system is calculated through inertial demand assessment.

[0013] The energy unit inertia capability assessment module is used to calculate the maximum virtual inertia that each energy unit can provide based on the physical model and operating constraints of each energy unit in the microgrid system.

[0014] The virtual inertia allocation optimization module is used to calculate the actual allocated virtual inertia of each energy unit based on the required virtual inertia and the maximum available virtual inertia, using grey relational analysis.

[0015] The additional power command generation module is used to generate additional power commands for each energy unit based on the actual allocated virtual inertia and the corresponding operation control model of each energy unit.

[0016] The collaborative power regulation execution module is used to superimpose the additional power commands corresponding to each energy unit onto the local control loop of the microgrid system, driving the power electronic converter to perform power regulation in order to carry out virtual inertial control of multi-energy collaboration.

[0017] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement a multi-energy coordinated smart microgrid control method as described in the first aspect.

[0018] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a multi-energy coordinated smart microgrid control method as described in the first aspect.

[0019] The aforementioned multi-energy collaborative smart microgrid control method, device, equipment, and medium accurately assess the system inertia demand by real-time acquisition of microgrid frequency signals and calculation of their rate of change. It then dynamically calculates the maximum virtual inertia each energy unit can provide, combining the physical characteristics and operational constraints of each unit. Using grey relational analysis, it dynamically matches system demand with unit capabilities to generate an optimal allocation scheme. This drives the dedicated control models of each unit to generate differentiated power regulation commands. Finally, power support is collaboratively executed through power electronic converters. This achieves precise matching and efficient collaboration of multi-energy virtual inertia capabilities, significantly enhancing the frequency stability of high-proportion renewable energy grids. It effectively avoids problems such as resource allocation mismatch and poor control adaptability in traditional methods, ensuring the dynamic safety of the system under complex operating conditions. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 A flowchart illustrating a multi-energy coordinated smart microgrid control method provided by the present invention;

[0022] Figure 2 This is a schematic diagram of the process of performing grey relational analysis in an optional embodiment of the present invention;

[0023] Figure 3 This is a schematic diagram of the structure of a multi-energy coordinated intelligent microgrid control device provided by the present invention. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0025] refer to Figure 1 The document presents a flowchart illustrating a multi-energy coordinated smart microgrid control method provided in this application, which includes the following steps:

[0026] S1. Frequency signals of the microgrid system are collected by synchronous phasor measurement units deployed at key nodes of the microgrid; based on the frequency change rate of the frequency signals, the required virtual inertia of the microgrid system is calculated through inertial demand assessment.

[0027] Specifically, the deployment of synchronous phasor measurement units (PMUs) covers key nodes of the microgrid, including the main power supply bus, grid connection points of various energy units, and important load access points. The selected PMUs comply with relevant industry standards for synchronous phasor measurement in power systems and possess voltage measurement accuracy, frequency measurement accuracy, and sampling rate sufficient to meet the frequency dynamic characteristics acquisition requirements of low-inertia microgrids. The PMUs are connected to each node via voltage transformers (TVs) and current transformers (TAs). The TV ratio is matched to the rated voltage of the corresponding node, and the TA ratio is matched to the rated current of the corresponding node to ensure the accuracy of the acquired signals. Signal transmission adopts methods compliant with power system communication specifications and follows relevant international or industry communication standards to ensure that communication delays are controlled within the range required for real-time frequency signal analysis.

[0028] The rate of change of frequency (df / dt) can be calculated using the five-point sliding window numerical differentiation method. The specific formula is as follows:

[0029]

[0030] Where T is the PMU sampling period, and f(t+nT) is the sampling frequency at time t+nT. This method effectively suppresses the interference of sampling noise, such as switching noise of power electronic equipment, on the calculation of the rate of change by weighted differentiation of the frequencies of five adjacent sampling points.

[0031] Inertial demand assessment can be performed using a two-factor weighted model of "frequency deviation - rate of change", the formula of which is:

[0032]

[0033] in, Virtual inertia required (unit: s). ( (System rated frequency, unit: Hz) , These are the weighting coefficients.

[0034] Coefficient calibration can be completed through offline simulation: A microgrid simulation model incorporating multiple energy sources such as wind power, photovoltaics, and energy storage is constructed to simulate typical system disturbance scenarios such as sudden load increases, sudden drops in renewable energy output, and off-grid disconnection of energy storage, recording different... , The maximum frequency deviation and frequency recovery time of the system under the combined conditions are selected. The optimal value is the combination of coefficients that makes the maximum frequency deviation of the system meet the grid frequency stability standard and the recovery time meet the dynamic response requirements under all disturbance scenarios. This ensures that the calculation of the required virtual inertia can adapt to the inertial replenishment needs of the system under different disturbances.

[0035] S2. Based on the physical model and operating constraints of each energy unit in the microgrid system, calculate the maximum virtual inertia that each energy unit can provide.

[0036] Specifically, the physical models of each energy unit are constructed based on its own topology and operating characteristics. For wind turbines, a third-order model of "wind speed-aerodynamic power-rotor speed-converter control" can be adopted, where the aerodynamic power formula can be... ( air density, This is the function for calculating the wind energy utilization coefficient. For the tip speed ratio, The pitch angle is the propeller angle. For swept area, (For wind speed), the rotor-side converter adopts a vector control strategy, which realizes independent control of active and reactive power by adjusting the d-axis and q-axis currents. Its virtual inertia provision capability is constrained by the rated capacity and overload capacity of the converter, and inertial output is performed within the allowable current and power range of the converter.

[0037] For photovoltaic arrays, a "light intensity-temperature-PV characteristic" model can be adopted. The array consists of several photovoltaic modules connected in a specific series and parallel manner to match the system voltage and power requirements. The inverter adopts a three-phase full-bridge topology and has maximum power point tracking (MPPT) function and virtual inertia control interface. MPPT maximizes the extraction of photovoltaic output through a specific algorithm. Its virtual inertia output is affected by the current light intensity. When the light intensity is lower than a certain threshold, the photovoltaic output is insufficient to support the virtual inertia output. At this time, the unit does not have the ability to provide virtual inertia.

[0038] For energy storage systems, a "SOC-voltage-charge / discharge current" model can be adopted. The system consists of several energy storage cells connected in series and parallel to meet voltage and capacity requirements. It is equipped with a bidirectional DC / DC converter to realize bidirectional energy flow. The operating constraints include the SOC operating range, charge / discharge current limit, and instantaneous power limit. The SOC is controlled within a reasonable range to avoid overcharging and over-discharging. The charge / discharge current and power are within the rated range allowed by the converter and energy storage cells, ensuring the safe and stable operation of the energy storage system while providing virtual inertia support.

[0039] The calculation of the maximum virtual inertia that each energy unit can provide is based on the "power-inertia" equivalence relationship, and the formula can be:

[0040]

[0041] in, For the i-th type of energy, the maximum virtual inertia that can be provided is This represents the current available power of the energy source. For the rated power of this energy source, This serves as the baseline inertia for the energy source (provided by the equipment manufacturer or obtained through experimental calibration). During calculation, the current available power is determined based on the real-time operating status of each energy unit. If the current available power of an energy unit is 0, its maximum virtual inertia is 0, and it will not participate in the subsequent virtual inertia allocation process, thus avoiding the waste of system resources caused by invalid allocation.

[0042] S3. Based on the demand virtual inertia and the maximum available virtual inertia, grey relational analysis is used to calculate the actual allocated virtual inertia of each energy unit.

[0043] Specifically, the implementation of Grey Relational Analysis (GRA) follows the process of "reference sequence determination - comparison sequence determination - data preprocessing - correlation coefficient calculation - correlation degree calculation". First, the reference sequence and comparison sequence are determined. The reference sequence can be composed of the required virtual inertia calculated in step S1, and the comparison sequence is the maximum available virtual inertia of each energy unit. If there is a difference in dimensions between the reference sequence and the comparison sequence, an "initialization" method is used to eliminate the influence of dimensions. The initialization formula is as follows: ,in The processed data; The original data represents the value at the k-th moment in the comparison sequence of the i-th energy unit. If the two have the same dimensions, no data preprocessing is required, and the data can proceed directly to the subsequent calculation stage.

[0044] The correlation coefficient can be calculated using Dunk's correlation formula:

[0045]

[0046] in, Let be the correlation coefficient of the i-th comparison sequence at time k; The resolution coefficient is used to balance the resolution of the correlation coefficient. It can be set to 0.5 to balance the resolution effect and the calculation stability. It is the minimum difference between two levels, that is, the minimum value among all the differences between the comparison sequence and the reference sequence; The maximum difference between the two levels is the largest difference among all the differences between the comparison sequence and the reference sequence. The calculation first solves for the difference between each comparison sequence and the reference sequence, then determines the minimum and maximum differences at both levels, and finally substitutes these values ​​into the formula to obtain the correlation coefficient for each comparison sequence.

[0047] correlation The correlation coefficient can be the average value; when only single-time data is involved, the correlation degree equals the correlation coefficient. The actual allocation of virtual inertia is calculated based on the "correlation degree ratio," using the following formula:

[0048]

[0049] Where n is the number of energy units participating in the allocation. The sum of the correlation of all participating energy units.

[0050] After allocation, the results are verified for rationality. First, it is verified that the allocated inertia of each unit does not exceed its maximum available virtual inertia to avoid operational failures caused by exceeding the unit's capacity. Second, it is verified that the sum of the allocated inertia of all units is consistent with the required virtual inertia to ensure that the system's required inertia is adequately replenished, and finally, a balanced allocation of inertia and efficient utilization of system resources are achieved.

[0051] S4. Based on the actual allocation of virtual inertia, generate additional power commands for each energy unit through the corresponding operation control model.

[0052] Specifically, the operation control model of each energy unit is designed in conjunction with its topology characteristics to ensure that the additional power command can be seamlessly compatible with the local control loop, so as to achieve accurate output of inertia.

[0053] For doubly-fed induction generator (DFIG) wind turbines, a model of "rotor-side converter vector control + virtual inertia-added control" can be adopted. The basic control logic of the rotor-side converter is to control reactive power through d-axis current and active power through q-axis current. The calculation of reactive power and active power is based on the d-axis and q-axis components of the stator voltage and rotor current, respectively. The virtual inertia-added power command is converted into the q-axis current increment, and the conversion formula can be obtained as follows: ( This represents the q-axis current increment. Additional power command for wind power (This refers to the d-axis component of the stator voltage). In specific scenarios such as low-voltage ride-through, the influence of the q-axis component of the stator voltage on current calculation can be ignored to simplify the control logic.

[0054] The calculation of the additional power command is based on the power-frequency relationship of the virtual inertia, as shown in the formula:

[0055]

[0056] in, Assigning virtual inertia to wind power, This refers to the rated apparent power of wind power. The frequency change rate is used. The calculated q-axis current increment is superimposed on the q-axis current reference value of the rotor-side converter. The current tracking control of the converter enables rapid adjustment of wind power output to meet the virtual inertia replenishment requirements.

[0057] For photovoltaic inverters, a "MPPT control + virtual inertia droop control" model can be adopted. MPPT tracks the maximum power point of the photovoltaic array using a specific algorithm and outputs a basic power command. The virtual inertia-based additional power command employs a dual control strategy of "inertia-damping," as shown in the formula:

[0058]

[0059] in, Assign virtual inertia to photovoltaics. The rated apparent power of photovoltaic power. The damping coefficient is... This represents the frequency deviation. The damping coefficient, determined through simulation calibration, is used to suppress system frequency oscillations and improve frequency stability. The total power command is the sum of the basic power command and the additional power command. It is necessary to ensure that the total power command is non-negative to prevent the inverter from shutting down due to abnormal power command. The total power command is converted into a d-axis current reference value through the inverter's current loop control, driving the inverter to output the corresponding current and realizing the virtual inertia output of the photovoltaic unit.

[0060] For energy storage systems, a model of "bidirectional DC / DC droop control + virtual inertia compensation control" can be adopted. The basic control for energy storage is droop control, and the basic power command is determined by the droop coefficient and frequency deviation. The formula for the virtual inertia-compensated power command is:

[0061]

[0062] in, Assign virtual inertia to energy storage. The rated apparent power of the energy storage is the total power command, which is the sum of the base power command and the additional power command. The command symbol represents the charging and discharging state of the energy storage system. The total power command is converted into the duty cycle command of the bidirectional DC / DC converter. By adjusting the duty cycle, the output voltage and current of the converter are controlled, thereby controlling the charging and discharging power of the energy storage system. Finally, the PWM controller outputs a drive signal to enable the DC / DC converter to perform the corresponding power regulation operation, completing the virtual inertia replenishment of the energy storage unit.

[0063] S5. The additional power commands corresponding to each energy unit are superimposed on the local control loop of the microgrid system to drive the power electronic converter to perform power regulation, so as to carry out virtual inertial control of multi-energy coordination.

[0064] Specifically, for the rotor-side converter of a doubly-fed induction generator (DFIG) wind turbine, a dual closed-loop control architecture of "current loop-speed loop" can be adopted. The current loop uses a PI controller, with the input signal being the deviation between the q-axis current reference value (the sum of the basic reference value and the additional current increment) and the actual current, and the output signal being the modulation signal of the rotor-side converter. The speed loop also uses a PI controller, with the input signal being the deviation between the speed reference value (determined by MPPT) and the actual speed, and the output signal being used to correct the q-axis current reference value of the current loop, forming a closed-loop control to improve the tracking accuracy of speed and current.

[0065] The converter can use IGBT modules that meet the power rating requirements. The drive circuit uses isolated drive chips to ensure the safety and anti-interference capability of the control signal. The drive signal is generated through space vector pulse width modulation (SVPWM) technology to ensure that the current tracking error is controlled within the allowable range. At the same time, the converter is equipped with overcurrent protection and overvoltage protection functions. The overcurrent protection sets the action threshold according to the rated current of the converter, and the overvoltage protection sets the action threshold according to the safe voltage of the DC bus. When the current or voltage exceeds the threshold, the protection mechanism is triggered quickly to avoid hardware damage caused by additional power commands.

[0066] For photovoltaic inverters, a dual closed-loop control architecture of "voltage loop-current loop" can be adopted. The voltage loop uses a PI controller, with the input signal being the deviation between the DC bus voltage reference value and the actual voltage, and the output signal being the d-axis current base reference value. The current loop also uses a PI controller, with the input signal being the deviation between the total d-axis current reference value (the sum of the base reference value and the additional reference value) and the actual current, and the output signal being the inverter's SVPWM modulation signal. The inverter uses IGBT modules that meet the power and voltage ratings, and sets a reasonable switching frequency to balance switching losses and output waveform quality. The filter circuit uses an LC filter structure to suppress output current ripple and ensure that the current waveform meets grid connection standards. In addition, the inverter also introduces a phase-locked loop (PLL), using synchronous rotating coordinate system PLL technology to track the grid voltage phase, ensuring that the phase of the additional power command output is consistent with the grid phase, avoiding reactive power fluctuations caused by phase deviation, and ensuring grid power quality.

[0067] For bidirectional DC / DC converters in energy storage systems, a "voltage loop-current loop" control architecture can be adopted. The voltage loop uses a PI controller, with the input signal being the deviation between the battery terminal voltage reference value and the actual voltage, and the output signal being the current reference value. The current loop also uses a PI controller, with the input signal being the deviation between the charging / discharging current reference value (calculated from the total power command) and the actual current, and the output signal being a PWM drive signal. The DC / DC converter employs a phase-shifted full-bridge topology to achieve wide-range voltage regulation and efficient energy conversion. IGBT modules matching the voltage and current ratings are selected, and an appropriate switching frequency is set. A filter circuit composed of inductors and capacitors is used to ensure that current ripple is controlled within the allowable range, guaranteeing the stable operation of the energy storage system.

[0068] Synchronization of multi-energy coordinated control is achieved through a central control unit (CCU). The CCU possesses sufficient computing power to meet real-time calculation requirements. It establishes communication connections with each power management unit (PMU) and energy unit controller via a bus conforming to power system communication standards. The CCU periodically receives frequency signals collected by each PMU and the operating status (including power, current, and SOC) of each energy unit. Based on the received signals, the CCU calculates the required virtual inertia and the allocated inertia for each unit, generating additional power commands for each energy unit and sending them to the corresponding unit controller via the communication bus. Upon receiving the commands, each unit controller updates its modulation signal within one control cycle, ensuring synchronized power regulation responses from all energy units and avoiding secondary frequency fluctuations caused by response delays that could affect system stability. Experimental verification shows that under this control process, the microgrid's frequency characteristics under typical disturbance scenarios meet stability requirements, and the output deviation of each energy unit is controlled within allowable limits, effectively achieving the virtual inertial control objective of multi-energy coordination.

[0069] The aforementioned multi-energy collaborative smart microgrid control method accurately assesses the system's inertia demand by real-time acquisition of microgrid frequency signals and calculation of their rate of change. It then dynamically calculates the maximum virtual inertia each energy unit can provide, combining the physical characteristics and operational constraints of each unit. Using grey relational analysis, it dynamically matches system demand with unit capabilities to generate an optimal allocation scheme. This drives the dedicated control models of each unit to generate differentiated power regulation commands. Finally, power support is collaboratively executed through power electronic converters. This achieves precise matching and efficient collaboration of multi-energy virtual inertia capabilities, significantly enhancing the frequency stability of high-proportion renewable energy grids. It effectively avoids problems such as resource allocation mismatch and poor control adaptability in traditional methods, ensuring the dynamic safety of the system under complex operating conditions.

[0070] In one optional embodiment, the energy unit includes a wind turbine, a photovoltaic system, and an energy storage system; based on the physical model and operating constraints of each energy unit in the microgrid system, the maximum virtual inertia that each energy unit can provide is calculated, including the following steps:

[0071] S11. Based on the rotor speed and output power of the wind turbine, the current releaseable rotational kinetic energy is calculated through the wind turbine dynamics model to obtain the current releaseable rotational kinetic energy of the wind turbine.

[0072] Specifically, the dynamic model of the wind turbine is based on the mechanical characteristics of the doubly-fed induction generator (DFIG), and the calculation of the rotor rotational kinetic energy is based on the formula. (Where J is the rotor moment of inertia, which is specified by the unit manufacturer;) Let n be the rotor angular velocity, which satisfies the following condition with respect to the rotational speed n. In actual operation, the rotor speed is acquired in real time through a shaft encoder, and the output power is acquired in combination with the voltage and current sensors of the converter. At the same time, considering the balance between aerodynamic power and mechanical power (aerodynamic power... Mechanical power ,in R is the air density, R is the rotor radius, and v is the wind speed. This is the function for calculating the wind energy utilization coefficient. (For mechanical friction loss), the rate of change of rotor kinetic energy satisfies Therefore, the "currently releaseable rotational kinetic energy" can be obtained by integrating the power difference over the most recent 10 seconds. Obtained, or obtained through "reference kinetic energy (kinetic energy at rated speed)". ")" and "current kinetic energy ( The difference between the current speed and the rated speed is used to accurately capture the real-time releaseable amount of kinetic energy of the wind turbine rotor. Specifically, when the current speed is lower than the rated speed, the difference is the releaseable kinetic energy.

[0073] S12. Based on the current releaseable rotational kinetic energy of the wind turbine and the preset maximum allowable speed deviation, calculate the maximum virtual inertia that the wind turbine can provide using the kinetic energy-inertia conversion formula; the kinetic energy-inertia conversion formula is:

[0074]

[0075] in, This provides the maximum virtual inertia for wind turbine units. The current available rotational kinetic energy of the wind turbine. This represents the current rotor angular velocity of the wind turbine. This is the preset maximum allowable speed deviation.

[0076] Specifically, the kinetic energy to inertia conversion formula The derivation is based on the physical nature of the frequency of the inertia-supported system. The inertia J of the synchronous generator can be obtained through... Correlation between power change and angular acceleration (approximately) When a wind turbine releases rotor kinetic energy to support the system's inertia, the rotor speed deviation must be limited to a safe range. Therefore, a preset maximum allowable speed deviation is introduced. It is determined by the constraints of safe operation of the unit.

[0077] During implementation, A 5% safety margin can be reserved (e.g., if the manufacturer specifies that the speed deviation exceeds 15% to trigger protection, the default is 10%). The value obtained from S11... Current rotor angular velocity , Substituting into the formula, we can obtain the maximum virtual inertia that the wind turbine can provide. For example, if , (Corresponding to a rotational speed of 1350 r / min) ,but This value matches the actual inertia of the unit, ensuring that the rotational speed does not exceed the safe range while maximizing the utilization of kinetic energy.

[0078] S13. Based on the illuminance, operating point voltage, and current data of the photovoltaic system, calculate the maximum available power of the photovoltaic system using the maximum power point tracking algorithm; subtract the current output power of the photovoltaic system from the maximum available power to obtain the power margin of the photovoltaic system.

[0079] Specifically, maximum power point tracking (MPPT) of a photovoltaic system can be achieved using the perturbation-observation method. The photovoltaic controller collects real-time data on light intensity, operating point voltage V, and current I through a light sensor, a voltage sensor (connected in parallel to the photovoltaic array output), and a current sensor (connected in series). This is combined with a single-diode model of the photovoltaic module ( ,in For photocurrent, Here, q represents the reverse saturation current, and q represents the electron charge. For series resistance, (where n is the ideality factor, k is the Boltzmann constant, and T is the temperature), and a temperature coefficient is introduced to correct for the effect of temperature on power. The maximum power point voltage is calculated under the current illumination and temperature. and current Thus, the maximum available power is obtained. Current output power It is obtained by multiplying V and I in real time. ), power margin for For example, when the light intensity is 800 W / m² and the temperature is 25°C, MPPT calculates... Current actual output ,but This margin provides the power basis for the virtual inertia of the photovoltaic system.

[0080] S14. Based on the power margin of the photovoltaic system and the preset maximum allowable frequency change rate of the microgrid system, calculate the maximum virtual inertia that the photovoltaic system can provide using the power margin-inertia mapping model; the power margin-inertia mapping model is as follows:

[0081]

[0082] in, This represents the maximum virtual inertia that a photovoltaic system can provide. For power margin, The angular velocity corresponding to the standard frequency of the microgrid system. This is the preset maximum allowable rate of frequency change.

[0083] Specifically, the power margin-inertia mapping model This stems from the relationship between virtual inertia and the rate of change of frequency (RoCoF). The inertia of a synchronous generator satisfies... , This is the rated angular frequency. The power margin of the photovoltaic system. This can be considered as additional power support capability, and therefore mapped to "equivalent virtual inertia" through this formula. The preset maximum allowable rate of frequency change... It must meet the stability requirements of microgrids (e.g., national standards stipulate that the grid RoCoF should not exceed 2Hz / s, and microgrids can be preset). Take the absolute value ). S13 Substitute into the formula, for example ,but This value represents the equivalent inertia that a photovoltaic system can simulate, and is used to support system frequency stability.

[0084] S15. Based on the state of charge, temperature and current charging and discharging current of the energy storage system, calculate the current available charging and discharging power limit of the energy storage system, and subtract the current actual power of the energy storage system from the current available charging and discharging power limit to obtain the charging and discharging power margin of the energy storage system.

[0085] Specifically, the usable charge / discharge power limit of an energy storage system (taking lithium batteries as an example) is constrained by three factors: state of charge (SOC), temperature, and charge / discharge current. SOC calculation can be performed using the ampere-hour integration method combined with open-circuit voltage calibration, as shown in the formula: ,in For the initial SOC, I represents the rated capacity, and I represents the charging / discharging current, with charging being positive and discharging being negative. Temperature constraints can limit the charging and discharging power using a pre-stored "power-temperature curve," such as charging power at 50% and discharging power at 80% of the rated capacity at 0℃; full power at 25℃; and charging and discharging at 60% of the rated capacity at 45℃. Current constraints are determined by the battery's rate characteristics; for example, at a rated rate of 1C, the charging current is limited to 0.2C when SOC ≥ 90%, and the discharging current is limited to 0.2C when SOC ≤ 10%.

[0086] Considering the above constraints, the current available charging power limit is determined by using a lookup table method in the battery management system (BMS) (pre-stored SOC-temperature-power matrix). and the current available discharge power limit Current actual power The voltage and current sensors of the energy storage converter are used to collect and calculate the data, and the formula is as follows: .

[0087] Therefore, charging power margin ,like For charging power, This refers to the remaining chargeable power; if it refers to the discharge power, Discharge power margin ,like For discharge power, This refers to the remaining discharge power; if it refers to the charging power, .

[0088] S16. Based on the charge / discharge power margin of the energy storage system and the preset maximum allowable frequency change rate of the microgrid system, calculate the maximum virtual inertia that the energy storage system can provide using a power-inertia linear conversion model; the power-inertia linear conversion model is as follows:

[0089]

[0090] in, The maximum virtual inertia that the energy storage system can provide. For the charging power margin of the energy storage system, This represents the discharge power margin of the energy storage system.

[0091] Specifically, the power-inertia linear conversion model The mapping logic is consistent with that of photovoltaic systems. The "sum of charge and discharge power margins" of an energy storage system is its total additional power capability, which is linearly converted into virtual inertia through its relationship with the rate of change of frequency and the rated angular frequency.

[0092] in the formula , Consistent with S14. Assuming energy storage charging power margin... Discharge power margin The total margin is 70kW. Substituting this into the equation, we get... This virtual inertia reflects the equivalent inertia provided by the energy storage system to the microgrid through charging and discharging regulation. Due to the fast response speed of energy storage, its dynamic inertia adjustment capability is better, and it can support the stability of system frequency more promptly.

[0093] In practice, the central controller needs to work with the local controllers of wind turbines, photovoltaics, and energy storage to collect data in real time and calculate the maximum virtual inertia that each unit can provide according to the above steps. This provides a basis for subsequent inertia allocation (such as grey relational analysis) and ensures the accurate implementation of virtual inertial control for multi-energy collaboration.

[0094] In this embodiment, the maximum virtual inertia supply potential of each wind turbine, photovoltaic system, and energy storage system is precisely quantified to provide a basis for multi-energy collaborative inertial control.

[0095] refer to Figure 2 In one optional embodiment, based on the required virtual inertia and the maximum available virtual inertia, grey relational analysis is used to calculate the actual allocated virtual inertia of each energy unit, including the following steps:

[0096] S21. Based on the maximum available virtual inertia and response characteristics of each energy unit, construct a comparison sequence for each energy unit; the comparison sequence is as follows: ;in, This represents the comparison sequence for the i-th energy unit; This represents the response characteristic index of the i-th energy unit; This represents the maximum virtual inertia that the i-th energy unit can provide.

[0097] Specifically, response characteristic indicators The response speed of an energy unit to the system's inertia requirements can be determined by combining the control architecture and physical characteristics of each energy unit. (Wind turbine unit) The control response delay for rotor kinetic energy release can be taken as the time difference between the detection of a frequency change by the synchronous phasor measurement unit (PMU) and the output of additional power by the wind turbine converter. This can be obtained through laboratory testing. (Photovoltaic system) The adjustment response time of the Maximum Power Point Tracking (MPPT) algorithm is taken, which is the cycle of "voltage disturbance - power detection - reference value update" in the disturbance observation method. Energy storage system The charge / discharge command response time of the bidirectional DC / DC converter is taken as the delay from when the battery management system (BMS) receives the additional power command to when the converter output power is adjusted.

[0098] When constructing the comparison sequence, each energy unit corresponds to a two-dimensional vector. ,in The "maximum available virtual inertia" is calculated in step S2 (or corresponding sub-steps S11-S16). For example, a certain wind turbine... , Then its comparison sequence is A certain energy storage system , Then the comparison sequence is By simultaneously incorporating "response speed" and "inertia capability," grey relational analysis can achieve a synergistic consideration of "fast response" and "large capability," avoiding resource waste or response delays caused by allocation based on a single dimension.

[0099] S22. Calculate the correlation coefficient between the comparison sequence of each energy unit and the preset ideal sequence using the grey relational coefficient formula, and use this coefficient as the grey relational coefficient for each energy unit; the grey relational coefficient formula is:

[0100]

[0101] in, The correlation coefficient of the i-th energy unit in the k-th dimension of the comparison sequence is represented by k=1 or 2. When k=1, it represents the response characteristic dimension, and when k=2, it represents the maximum available inertia dimension. This represents the characteristic value of the k-th dimension of the ideal sequence. This represents the characteristic value of the k-th dimension of the comparison sequence of the i-th energy unit; , ρ is the preset resolution coefficient.

[0102] Specifically, constructing ideal sequences It is necessary to reflect the optimal characteristics of each dimension. For the response characteristic dimension (k=1), the minimum response time among all energy units is selected (i.e., (), representing the ideal characteristic of "fastest response"; for the maximum available inertia dimension (k=2), the largest among all energy units is selected. (Right now This represents the ideal characteristic of "strongest inertia capability." It includes wind turbine units (…). ), photovoltaic ( ), energy storage ( Taking a microgrid as an example, the ideal sequence is: .

[0103] When calculating the grey relational coefficient, first determine the minimum difference between the two levels. The maximum difference between the two levels , This means iterating through both dimensions of all energy units, calculating the absolute difference from the ideal sequence, and taking the minimum value. Similarly, take the maximum value of the absolute difference.

[0104] For the example above, the absolute difference between each unit and the ideal sequence is:

[0105] Wind turbine: |3-8|=5 (k=1), |225-225|=0 (k=2);

[0106] Photovoltaics: |3-50|=47 (k=1), |225-42|=183 (k=2);

[0107] Energy storage: |3-3|=0 (k=1), |225-148|=77 (k=2);

[0108] therefore, , Resolution coefficient It can be preset to 0.5 to balance the resolution and anti-interference ability of the correlation coefficient. Substitute it into the grey relational coefficient formula. Calculate the correlation coefficients of each unit in two dimensions:

[0109] Wind turbine ( ): ; ;

[0110] Photovoltaics ( ): ; ;

[0111] Energy storage ( ): ; .

[0112] The larger the correlation coefficient, the closer the unit is to the ideal characteristics in the corresponding dimension.

[0113] S23. Based on the grey relational coefficients of each energy unit, calculate the grey relational degree of each energy unit using the arithmetic mean algorithm; the formula for calculating the grey relational degree is:

[0114]

[0115] in, Let represent the grey relational degree of the i-th energy unit, and N be the number of feature indicators of the comparison sequence.

[0116] Specifically, grey relational degree It is the arithmetic mean of the correlation coefficients across all dimensions, used to comprehensively evaluate the overall similarity between the energy unit and the ideal sequence. The formula is: , where N is the number of feature indicators of the comparison sequence (here N=2, including two dimensions: "response characteristics" and "maximum available inertia").

[0117] Based on the calculation results of S22, the grey relational degree of each unit is:

[0118] Wind turbine units: ;

[0119] Photovoltaics: ;

[0120] Energy storage: .

[0121] It is evident that wind turbines have the highest correlation, indicating that their overall performance in terms of "response speed" and "inertia capability" is closest to the ideal state; photovoltaics have the lowest correlation due to their slow response speed and weak inertia capability; although energy storage has the fastest response, its inertia capability is inferior to that of wind power, so its correlation is in the middle.

[0122] S24. Based on the grey relational degree and maximum available virtual inertia of each energy unit, calculate the allocation coefficient of each energy unit using the proportional allocation formula; the proportional allocation formula is:

[0123]

[0124] in, This represents the allocation coefficient for the i-th energy unit.

[0125] Specifically, allocation coefficient It also reflects the "matching degree with the ideal sequence (grey relational degree)". ")" and "Maximum inertia capability ()" The formula is: The numerator is the product of "correlation degree - inertia capability", and the denominator is the sum of this product of all units, ensuring that the allocation coefficient is positively correlated with units that have "high matching degree and strong capability".

[0126] Based on the data above (wind turbines) , Photovoltaics , Energy storage , Calculate the numerator:

[0127] Wind turbine units: ;

[0128] Photovoltaics: ;

[0129] Energy storage: .

[0130] The denominator is the sum of the three, that is... .

[0131] Therefore, the allocation coefficient for each unit is:

[0132] Wind turbine units: ;

[0133] Photovoltaics: ;

[0134] Energy storage: .

[0135] The allocation coefficient reflects the proportion that each unit should bear in the total inertia demand. Wind turbines have the highest allocation ratio because of their best overall performance, while photovoltaics have the lowest allocation ratio because of their worst overall performance.

[0136] S25. Based on the allocation coefficient and the virtual inertia of demand, calculate the actual allocated virtual inertia of each energy unit using a linear allocation formula; the linear allocation formula is:

[0137]

[0138] in, The actual virtual inertia is allocated to the i-th energy unit. For virtual inertia required.

[0139] Specifically, the actual allocation of virtual inertia Through "allocation coefficient" "and demand virtual inertia" The linear product of ' is obtained, .in It is the virtual inertia required by the microgrid system calculated in step S1.

[0140] like Based on the allocation coefficient of S24, the actual virtual inertia allocated to each unit is:

[0141] Wind turbine units: ;

[0142] Photovoltaics: ;

[0143] Energy storage: .

[0144] This allocation ensures that the total inertia matches the demand. This ensures that the inertia allocated to each unit does not exceed its "maximum available virtual inertia" (185.7≤225, 17.7≤42, 96.6≤148), and ensures that "response speed" and "inertia capacity" are comprehensively considered, so that inertia replenishment is both efficient and timely, and avoids increased system frequency fluctuations caused by overload or response lag of a single unit.

[0145] In this embodiment, the response characteristics and maximum inertia capability of the integrated energy unit are combined to achieve optimal allocation of virtual inertia, thereby improving the rationality and synergy of inertia allocation.

[0146] In one optional embodiment, the step of generating the additional power command corresponding to the wind turbine includes:

[0147] S31. Based on the actual allocated virtual inertia of the wind turbine, the basic additional power command of the wind turbine is generated by combining the synchronous generator rotor motion equation model with real-time frequency change rate and frequency deviation data.

[0148] Specifically, the synchronous generator rotor motion equation is the core theoretical basis for generating the additional power command for wind turbine foundations, and its classic form is: Where J is the moment of inertia of the synchronous generator rotor. For rotor electrical angle, For mechanical torque, This represents the electromagnetic torque. To adapt to the virtual inertia control scenario of wind turbine units, this equation is transformed into a relationship between power and frequency, combined with the relationship between power and torque. ( The rotor angular velocity is related to the system frequency. satisfy ), and the conversion relationship between the virtual inertia time constant H and the rotational inertia J. ( The rated angular velocity, (Given the rated apparent power of the wind turbine), the power-frequency dynamic equation, which includes virtual inertia, can be derived:

[0149]

[0150] in For frequency deviation, For mechanical power, This refers to electromagnetic power.

[0151] In the virtual inertia control of wind turbines, the basic additional power command is essentially achieved by adjusting the electromagnetic power. The inertial response of a synchronous generator is simulated to compensate for system frequency fluctuations. This is combined with the virtual inertia of the wind turbine generator obtained in step S3. Replace H in the above equation with H and introduce the real-time frequency change rate. Frequency deviation As input variables, the calculation formula for the basic additional power command can be further derived:

[0152]

[0153] In the formula Add power commands to the base. The damping coefficient is used to suppress frequency oscillations. Its value can be calibrated through offline simulation. It can usually be determined based on the system damping ratio requirement (generally 0.4~0.6) and the closed-loop bandwidth (usually 1~5Hz) to ensure that the dynamic response speed of the system is not affected while suppressing oscillations.

[0154] Real-time frequency change rate Frequency deviation The frequency deviation can be obtained from real-time monitoring data of the microgrid. The frequency deviation is calculated directly from the difference between the real-time frequency collected by the PMU deployed at the wind turbine grid connection point and the system's rated frequency. The frequency change rate is processed using the five-point sliding window numerical differentiation method, consistent with step S1, to avoid interference from power electronic device switching noise on the calculation accuracy. In practical applications, it is necessary to ensure that the PMU data sampling period matches the wind turbine controller's operation period, so that the basic additional power command can follow the dynamic changes in system frequency in real time and accurately respond to inertial compensation requirements.

[0155] S32. Based on the basic additional power command of the wind turbine and the rotor speed change trend, the basic additional power command is dynamically corrected when a speed recovery requirement is detected through the speed recovery limiting mechanism, so as to obtain the corrected additional power command of the wind turbine.

[0156] Specifically, the design purpose of the speed recovery limiting mechanism is to prevent the rotor speed of wind turbines from dropping excessively due to the continuous provision of virtual inertia, thereby affecting the safe operation of the unit and its subsequent power generation capacity. The core of this mechanism is to determine whether the basic additional power command needs to be corrected based on the rotor speed change trend. The specific implementation is divided into two stages: "speed change trend detection" and "dynamic correction logic execution".

[0157] Rotor speed change trend detection is achieved by acquiring real-time rotor speed data from the wind turbine shaft encoder. A moving average method can be used to process the continuous speed signal to reduce measurement noise, and then the rate of change of speed can be calculated. (n is the rotor speed) is used to determine the trend. When If the duration exceeds the set threshold, or the real-time speed is lower than the rated speed by a certain percentage (the specific value can be determined in conjunction with the safe operating range provided by the unit manufacturer), it is determined that there is a need for speed recovery, triggering dynamic correction; otherwise, if the speed is maintained within the rated range and the rate of change is close to 0, no correction is required, and the basic additional power command is directly used.

[0158] The dynamic correction logic adopts a "speed deviation-correction coefficient" correlation model, introducing a correction coefficient. The base additional power command is scaled, and the corrected additional power command formula is as follows:

[0159]

[0160] Where the correction coefficient Based on speed deviation ( The rated speed is designed as a piecewise function. When the speed is higher than or equal to the rated value, No correction; when ( When the first threshold is 5% of the rated speed, it can be calculated using the formula... Linearly reduce the correction factor to slow down the speed drop; when At that time, a formula can be used. ( The second threshold (which can be 10% of the rated speed) further reduces the correction factor to prioritize speed recovery; when When, it can be defined as Limit the output of the basic additional power command to prevent the speed from continuing to drop below the safe threshold.

[0161] A hysteresis loop can be introduced during the correction process to prevent frequent switching of the correction coefficient due to minor fluctuations in rotational speed, which could lead to oscillations in the additional power command and affect the stability of the converter operation. Simultaneously, the corrected additional power command is compared with the current available power of the wind turbine to ensure that the corrected additional power command does not exceed the current available power (determined by the calculation result of "current available power of the wind turbine" in step S2), thus avoiding overcurrent protection activation of the converter due to command overload.

[0162] S33. By superimposing the corrected additional power command onto the original power reference value through the power control loop of the wind turbine rotor-side converter, the final active power regulation command of the wind turbine is obtained.

[0163] Specifically, the power control loop of the wind turbine rotor-side converter adopts a dual closed-loop architecture of "power loop-current loop". This architecture can achieve accurate tracking of active power while ensuring the stability of converter operation. The power loop is the outer loop and the current loop is the inner loop. The correction command is transmitted through the hierarchy of this dual closed loop and is finally converted into the modulation signal of the converter.

[0164] First, correct the additional power command. This is superimposed on the original power reference value of the wind turbine. The original power reference value is generated by the Maximum Wind Energy Tracking (MPPT) algorithm, the core of which is to adjust the rotor speed according to the real-time wind speed to capture the maximum wind energy and output an active power reference value that matches the current wind speed. The reference value for the total active power after superposition is... This total reference value serves as the input to the power loop and is correlated with the actual output active power of the converter. Forming deviation signal .

[0165] The power loop uses a PI controller to control the deviation signal. Adjustments are made to output the q-axis current reference value for the current loop. This transformation is based on the stator flux orientation vector control principle of doubly-fed induction generator (DFIG) wind turbines. In the stator flux orientation coordinate system, active power control is achieved by adjusting the q-axis component of the rotor current, and the two satisfy a linear relationship. ( For mutual inductance between stator and rotor, (For the stator inductance), therefore the output of the power loop PI controller can be directly mapped to the q-axis current reference value, ensuring accurate correspondence between active power and q-axis current.

[0166] The current loop serves as the inner loop, and its input is the q-axis current reference value. The actual q-axis current collected by the rotor-side current sensor deviation Similarly, a PI controller is used for regulation. The output of the current loop PI controller is the modulation voltage reference value of the rotor-side converter. This reference value is converted into the IGBT drive signal through space vector pulse width modulation (SVPWM) technology, which controls the switching state of the converter, thereby adjusting the q-axis component of the rotor current, and finally achieving tracking of the total active power reference value.

[0167] In the parameter design of the power control loop, the principle of "inner loop bandwidth higher than outer loop" is followed. The current loop bandwidth can be set to 100~200Hz to quickly suppress current disturbances; the power loop bandwidth can be set to 10~20Hz to balance power tracking speed and system stability. Simultaneously, a feedforward compensation circuit is introduced, substituting real-time parameters such as stator flux linkage and rotor angular velocity into the current loop control equation to offset the impact of system coupling terms on current tracking accuracy. Furthermore, the converter is equipped with overcurrent and overvoltage protection logic. When the rotor current exceeds 1.2~1.5 times the rated value or the DC bus voltage exceeds 1.1~1.2 times the rated value, the active power reference value is immediately reduced to avoid hardware damage and ensure the safe operation of the wind turbine when executing active power regulation commands.

[0168] In this embodiment, the wind turbine provides virtual inertia while taking into account both the safe recovery of rotor speed and the inertial support requirements, ensuring stable and effective response of the unit.

[0169] In one optional embodiment, the step of generating the additional power command corresponding to the photovoltaic system includes:

[0170] S41. Based on the actual allocation of virtual inertia of the photovoltaic system, the basic additional power command of the photovoltaic system is generated through the virtual inertia-power conversion model.

[0171] Specifically, the virtual inertia-power conversion model for photovoltaic systems needs to be adapted to their characteristic of having no mechanically rotating parts. The core is to simulate the inertial response of a synchronous generator through power regulation, essentially converting the actually allocated virtual inertia into active power commands that can directly drive the inverter. The derivation of this model is based on the dynamic relationship between virtual inertia and frequency. The core function of virtual inertia is to suppress frequency changes by supplementing or absorbing active power. Its dynamic relationship with the rate of frequency change can be simplified and derived from the synchronous generator rotor motion equations. Considering the characteristic of photovoltaic systems that achieve power regulation solely through power electronic conversion, the virtual inertia-power conversion formula is finally obtained:

[0172]

[0173] In the formula Add power commands to the base of photovoltaic systems. The virtual inertia is actually allocated to the photovoltaic system (derived from the inertia allocation results of the grey relational analysis mentioned above). The rated apparent power of the photovoltaic system, This represents the real-time frequency change rate of the microgrid.

[0174] From a technical perspective, the physical meaning of this formula is: a photovoltaic system needs to output active power that is positively correlated with its virtual inertia, rated power, and frequency change rate to offset rapid fluctuations in system frequency. When the system frequency decreases ( )hour, A negative value indicates that the photovoltaic system needs to increase its active power output to support the frequency; when the system frequency increases ( )hour, A positive value indicates that the photovoltaic system needs to reduce its active power output (or absorbed power) to suppress excessively high frequencies.

[0175] Acquisition of model input parameters and coordination with the microgrid monitoring system: The central controller issues in real time based on the system's inertia requirements, and the update cycle is consistent with the inertia allocation cycle. The five-point sliding window numerical differentiation method, consistent with that used for wind turbines, is employed. Data is sourced from the PMU deployed at the grid connection point of the photovoltaic inverter, ensuring the accuracy of the frequency change rate calculation and avoiding command fluctuations caused by power electronic switching noise. Furthermore, the calculated... A limiting process is implemented, initially restricting its absolute value to no more than 30% to 50% of the current available power of the photovoltaic system (the specific percentage can be determined based on the stability of photovoltaic output), leaving room for adjustment and subsequent power reserve.

[0176] S42. Based on historical frequency disturbance statistics and future solar intensity predictions, the power reserve margin required by the photovoltaic system is calculated using a power margin adaptive algorithm.

[0177] Specifically, the design goal of the power margin adaptive algorithm is to ensure that the photovoltaic system, while providing virtual inertia support, reserves sufficient power to cope with potential future frequency disturbances and changes in irradiance, avoiding interruptions in inertial response due to insufficient power. This algorithm uses "historical disturbance demand - future output prediction" as dual inputs, and dynamically calculates the currently required power reserve margin through the synergy of statistical analysis and prediction models. The specific implementation is divided into three stages: "historical data statistics", "illuminance prediction modeling" and "margin dynamic calculation".

[0178] Historical frequency disturbances can be statistically analyzed by processing microgrid frequency disturbance data from the past 30 to 90 days. First, disturbance events with an absolute frequency deviation exceeding 0.1 Hz or an absolute frequency change rate exceeding 0.2 Hz / s can be selected. The maximum inertial power that the photovoltaic system needs to replenish during each event (i.e., the power required for that event) is recorded. (Peak values); then, these peak values ​​are fitted using a Weibull distribution or a normal distribution to calculate the power value corresponding to the 95% confidence interval. This value represents the maximum disturbance demand that is most likely to occur in historical data, and serves as a basic reference for power reserve margin.

[0179] Future solar irradiance forecasts employ short-term prediction models (prediction timeframes of 5-15 minutes, matching the photovoltaic power output response delay). Models used include time series analysis models (such as ARIMA), machine learning models (such as LSTM), or physical models (based on meteorological satellite cloud images and radiative transfer equations). The forecast output is the average solar irradiance over the future time period. Combined with the power-light characteristic curve of the photovoltaic array ( Where G is the light intensity, (Based on the light intensity under standard test conditions), calculate the predicted maximum available photovoltaic power for the next 5-15 minutes. Simultaneously, calculate the current maximum available power based on the current light intensity. The power fluctuation caused by changes in illumination was obtained. This fluctuation is incorporated into the power reserve margin to avoid insufficient photovoltaic output due to a sudden drop in sunlight.

[0180] The dynamic calculation of power reserve margin adopts a weighted summation model, and the formula is:

[0181]

[0182] In the formula , These are weighting coefficients, summing to 1, and their values ​​are adaptively adjusted based on the microgrid's operating scenario. For example, when historical disturbances are frequent (e.g., the average number of disturbances per day exceeds 5), Take a value of 0.6 to 0.7. Use a value of 0.3~0.4 to prioritize meeting historical disturbance needs; when there are drastic changes in illumination (such as cloudy weather, with illumination fluctuations exceeding 200W / m² within 10 minutes), Take a value of 0.4~0.5. A margin of 0.5-0.6 is preferred to address fluctuations in sunlight. Furthermore, upper and lower margins are set: the lower margin can be 10% of the current maximum available power (to avoid an excessively small margin rendering the reserve meaningless), and the upper margin can be 40% of the current maximum available power (to avoid an excessively large margin wasting photovoltaic output), ensuring... Adjust dynamically within a reasonable range.

[0183] S43. Based on the basic additional power command and power reserve margin of the photovoltaic system, the final active power regulation command of the photovoltaic system is generated through superposition calculation.

[0184] Specifically, the generation of the final active power regulation command of the photovoltaic system is achieved through the coordinated constraint of "basic additional power command - power reserve margin". The core is to ensure sufficient reserved power while meeting the inertial support requirements, so as to avoid the system falling into secondary disturbances due to insufficient power. This process, combined with the control architecture of the photovoltaic inverter, accurately transmits the command to the power control loop. The specific implementation includes three stages: "command superposition constraint", "integration with the original power reference value", and "control loop adaptation".

[0185] The core of the command superposition constraint is to ensure that the basic additional power command does not exceed the range of "current available power - power reserve margin". First, calculate the maximum power that the photovoltaic system can currently use for inertial support. ,in The maximum available power is calculated for the current MPPT algorithm; then additional power commands are applied to the base. Perform secondary amplitude limiting, when When (increased output is required), its lower limit is limited to: That is, the maximum increase in output should not exceed the power available for inertial support, so as to avoid exceeding the current output capacity of photovoltaic power; when When output needs to be reduced, its upper limit is limited to: To prevent the inverter from shutting down due to low output, among other things... This is the minimum operating power of the photovoltaic inverter.

[0186] The integration with the original power reference value is based on the control logic of the photovoltaic inverter. Original power reference value Generated by the MPPT algorithm, representing the maximum power output reference under the current illumination; the final active power adjustment command. The sum of the original reference value and the base additional power command after limiting is given by the following formula:

[0187]

[0188] In the formula This is the base additional power command after secondary limiting. The physical meaning of this fusion logic is: based on the maximum output of MPPT, the actual output is adjusted according to the system inertia requirements, that is, the output is increased when the frequency needs to be supported (by superimposing the negative base command, because...). hour Note the symbolic logic here: when the system frequency decreases, the photovoltaic output needs to increase. If the value is negative, the actual output after superposition is higher than the MPPT reference value. It is necessary to ensure that this value does not exceed [the reference value]. When it is necessary to suppress the rise in frequency, reduce the output (superimpose a positive base instruction, and the actual output is lower than the MPPT reference value).

[0189] The control loop is adapted to the dual closed-loop architecture of "voltage loop-current loop" for photovoltaic inverters. Final instructions. First, input the power loop and compare it with the actual active power output of the inverter. Formation of deviation The power loop uses a PI controller to adjust this deviation, and the output serves as the d-axis current reference value for the current loop. The conversion is based on the power-current relationship of the photovoltaic inverter. , (This refers to the d-axis component of the grid voltage, based on grid voltage-oriented vector control). The current loop compares the d-axis current reference value with the actual d-axis current collected by the current sensor, outputs the modulated voltage reference value through the PI controller, and then converts it into an IGBT drive signal through SVPWM technology to realize the execution of the final command.

[0190] In addition, a command smoothing mechanism (e.g., using a first-order low-pass filter) can be introduced to prevent current loop oscillations caused by rapid changes in the basic additional power command; simultaneously, overcurrent and overvoltage protection should be configured: for example, when the inverter output current is detected to exceed 1.2 to 1.5 times the rated value or the DC bus voltage exceeds 1.1 to 1.2 times the rated value, it should be immediately switched off. Down to This ensures the safe operation of the inverter and sends a fault signal to the central controller, triggering inertial support compensation from other energy units to maintain the stability of the microgrid frequency.

[0191] In this embodiment, the photovoltaic system provides inertial support while reserving power to cope with future frequency disturbances and changes in illumination, thus balancing inertial response and output stability.

[0192] In one optional embodiment, the step of generating the additional power command corresponding to the energy storage system includes:

[0193] S51. Allocate the actual virtual inertia of the energy storage system into fast response inertia and continuous support inertia according to a preset ratio.

[0194] Specifically, the proportion of virtual inertia actually allocated to the energy storage system should be adapted to its dual characteristics of "rapid response + continuous support," and the preset proportion can be determined based on the dynamic frequency characteristics of the microgrid. The actual allocation of virtual inertia can be... Divide into fast response inertia according to 6:4 or 7:3 With continuous support inertia ,Right now , , where k is the proportionality coefficient.

[0195] The core basis for the proportional division is the time-scale characteristics of frequency disturbances. The fast-response inertia targets the high-frequency rate of change (RoCoF) in the initial stage of frequency disturbance (0~2 seconds) by suppressing the rapid drop or rise of frequency through instantaneous power adjustment; the continuous support inertia targets the steady-state frequency deviation in the middle stage of disturbance (2~10 seconds) by maintaining the frequency within the allowable range through stable power output.

[0196] In practical applications, the proportional gain can be dynamically adjusted through offline simulation. Under typical disturbance scenarios such as sudden load increases and sharp drops in renewable energy output, the maximum frequency deviation and recovery time under different proportional gains are simulated, and the optimal proportional gain value (e.g., maximum deviation ≤ 0.5Hz, recovery time ≤ 10 seconds) is selected. For example, in microgrids with a high proportion of wind power, there are many high-frequency disturbances caused by wind power fluctuations; k can be set to 0.7 to enhance rapid response capabilities. In microgrids dominated by stable loads, k can be set to 0.6 to optimize continuous support.

[0197] S52. Based on the fast response inertia, an additional power command for the fast response stage is generated through the differential control loop.

[0198] Specifically, the generation of additional power commands during the fast response phase is based on the principle of differential control. Its core is to achieve instantaneous power compensation by capturing the rate of change of frequency (RoCoF), thus suppressing rapid dynamic frequency shifts. The input to the differential control loop is the real-time rate of change of frequency. Output and fast response inertia Direct association, the formula is:

[0199]

[0200] In the formula Additional power commands are added for the rapid response phase. The rated apparent power of the energy storage system is given by the coefficient 2, which is derived from the conversion relationship between virtual inertia and rotational inertia (consistent with the equation of motion of the synchronous generator rotor).

[0201] The physical meaning of this instruction is: when the system frequency drops rapidly ( )hour, A negative value indicates that the energy storage system needs to release power instantaneously (discharge) to support the frequency; when the frequency rises rapidly ( )hour, A positive value indicates that the energy storage system needs to absorb power instantaneously (charge) to suppress excessively high frequencies. To avoid command oscillations caused by high-frequency noise, a first-order low-pass filter can be added before the differential control stage to smooth the original frequency signal; at the same time, a command change rate limit is set to ensure that the power switching speed of the energy storage converter does not exceed its hardware capabilities (such as the switching frequency and current change rate limits of the IGBT), preventing overcurrent protection from activating.

[0202] S53. Based on the continuous support inertia, an additional power command for the continuous support stage is generated through the proportional control link.

[0203] Specifically, the generation of additional power commands during the continuous support phase is based on the proportional control principle. Its core is to offset steady-state frequency deviations with stable power output, maintaining the frequency near the rated value. The input to the proportional control loop is the real-time frequency deviation. ( (at rated frequency), output and continuous support inertia The relationship between the damping coefficient and the damping coefficient is shown in the following formula:

[0204]

[0205] In the formula Additional power commands will be provided for the sustained support phase. It is a time constant used to adjust the response speed of proportional control, balancing steady-state accuracy and dynamic stability.

[0206] The physical meaning of this instruction is: when the system has a negative frequency deviation ( When the frequency is lower than the rated value, A negative value indicates that the energy storage system needs to continuously discharge to replenish power; when there is a positive frequency deviation ( )hour, A positive value indicates that the energy storage system needs to continuously charge to absorb excess power.

[0207] The proportional control parameters are tuned using the root locus method or frequency domain analysis. Time constant. Too small a damping ratio will result in a large overshoot, while too large a ratio will lead to a slow response. The optimal damping ratio is to maintain the closed-loop system within the range of 0.5 to 0.7, ensuring that the time for the frequency deviation to converge to within ±0.1Hz does not exceed 10 seconds. Furthermore, for… Set steady-state limits to avoid long-term high-power charging and discharging, which can lead to a decrease in the lifespan of energy storage batteries.

[0208] S54. The additional power command for the rapid response phase and the additional power command for the continuous support phase are superimposed and calculated to obtain the total additional power command for the energy storage system.

[0209] Specifically, the superposition logic of the total additional power command reflects the energy storage system's "transient-steady-state" all-time response capability to frequency disturbances. Through the synergy of rapid response and continuous support, it achieves overall optimization of frequency dynamic characteristics. The superposition formula is:

[0210]

[0211] In the formula This is the total additional power command for the energy storage system. Its sign and magnitude reflect the charging and discharging state and intensity of the energy storage system.

[0212] when When the system is in discharge mode, it injects power into the grid; when... At this time, the energy storage system operates in charging mode, absorbing power from the grid; the larger the amplitude, the higher the charging and discharging intensity, and the more significant the frequency regulation effect.

[0213] The superposition process satisfies the "dynamic coordination constraint." The decay rate of the fast response command matches the establishment rate of the continuous support command, preventing them from superimposing in reverse during the transition phase (e.g., 2-3 seconds after a disturbance), which would cause fluctuations in total power. This is specifically achieved by introducing an exponential decay factor into the fast response command. The modified fast command is: ( (This is the decay time constant) to ensure that after the continuous support command is effectively established, the response command is quickly withdrawn, reducing energy loss.

[0214] S55. Based on the total additional power command and real-time state of charge of the energy storage system, the amplitude of the total additional power command is constrained through a dynamic limiting algorithm to obtain the final active power regulation command of the energy storage system.

[0215] Specifically, the core of the dynamic limiting algorithm is to adjust the amplitude of the total additional power command based on the real-time state of charge (SOC) of the energy storage system, avoiding battery life degradation or safety risks caused by overcharging and over-discharging, while maximizing the utilization of energy storage capacity in frequency regulation. This algorithm dynamically generates the limiting coefficient through a piecewise SOC function. By constraining the total additional power command, we finally obtain:

[0216]

[0217] where is the final active power regulation command of the energy storage system, and the piecewise logic of the amplitude limiting coefficient can be as follows:

[0218] SOC≥90% (high state of charge): During discharging ( ), (full power discharging is allowed, and frequency support is prioritized); during charging ( ), (charging is strictly restricted to avoid overcharging).

[0219] 70% < SOC < 90% (medium-high state of charge): During discharging, ; during charging, (charging is moderately restricted).

[0220] 30% ≤ SOC ≤ 70% (normal state of charge): During both charging and discharging, is taken (unrestricted, fully participating in regulation).

[0221] 10% < SOC < 30% (medium-low state of charge): During discharging, (discharging is moderately restricted); during charging, (full power charging is allowed).

[0222] SOC ≤ 10% (low state of charge): During discharging, (discharging is strictly restricted to avoid over-discharging); during charging, .

[0223] The specific values of the amplitude limiting coefficient can be calibrated according to the charging and discharging characteristics of battery types (such as lithium batteries and lead-acid batteries). For example, the lower limit of SOC for lithium iron phosphate batteries can be relaxed to 5%, while that for ternary lithium batteries needs to be strictly controlled above 10%. In addition, a hysteresis link of 1 - 2 seconds can be introduced during the amplitude limiting process to prevent the amplitude limiting coefficient from frequently switching due to the fluctuation of SOC near the threshold, which may cause power command oscillation. The final command is input into the power control loop of the energy storage bidirectional converter and is converted into a PWM drive signal through the double closed-loop regulation of "voltage loop - current loop" to achieve precise charging and discharging power control, ensuring the maximum frequency support ability of the energy storage system under safety constraints.

[0224] In this embodiment, by leveraging the characteristics of "fast response + continuous support" of the energy storage and combining with the dynamic constraints of the state of charge, inertial support is provided for the system efficiently and safely.

[0225] In one optional embodiment, the required virtual inertia of the microgrid system is calculated based on the frequency change rate of the frequency signal through inertial demand assessment, including the following steps:

[0226] S61. Based on the frequency signal, the frequency change rate is calculated using the center difference method to obtain the real-time frequency change rate; the formula for calculating the real-time frequency change rate is:

[0227]

[0228] Where RoCoF is the real-time frequency change rate. This represents the frequency value at time t+Δt. This represents the frequency value at time t-Δt, where Δt is the sampling interval.

[0229] Specifically, the core principle of the central difference method is to use the frequency values ​​of symmetrical sampling points before and after the current moment to derive the rate of frequency change through linear approximation, thereby reducing calculation errors. The frequency signal is synchronously sampled for voltage and current by synchronous phasor measurement units (PMUs) deployed at key nodes of the microgrid. The sampling frequency can be set to 100Hz (i.e., sampling interval). This ensures that the acquisition accuracy of frequency dynamic characteristics can match the rapid frequency changes under the low inertia of the microgrid.

[0230] The formula for calculating the real-time frequency change rate is: ,in The frequency measurement value at time t+Δt. This is the frequency measurement at time t-Δt. The formula originates from a first-order approximation of the Taylor expansion, relating the frequency function... When expanded at time t, the central difference can cancel out even-order error terms, resulting in higher accuracy and smaller phase delay compared to forward / backward differences.

[0231] To avoid measurement noise interference, the original frequency signal can be low-pass filtered before the center differential. The filter cutoff frequency can be set to 10Hz (higher than the highest frequency of the system's dynamic frequency change, while filtering out high-frequency noise from power electronic switches, etc.). A 32nd-order finite impulse response (FIR) filter can be used to balance the filtering effect with the computational load. For example, if the frequency at time t-Δt is 49.95 Hz and at time t+Δt is 49.85 Hz, then... The negative sign indicates that the frequency is decreasing.

[0232] S62. Identify the disturbance type identifier based on the real-time frequency change rate and the corresponding duration;

[0233] Among them, when And duration When, it is determined to be a unit disconnection disturbance; when and When this occurs, it is determined to be a sudden load disturbance; among which, To quickly set the perturbation threshold, The slow disturbance threshold, For short-term disturbance windows, This is a long-term disturbance window.

[0234] Specifically, the purpose of disturbance type identification is to distinguish between "unit disconnection disturbance" and "load change disturbance" based on the amplitude and duration of the frequency change rate, so as to provide a "disturbance intensity-type" correlation basis for the subsequent weighted assessment of demand virtual inertia.

[0235] Each threshold parameter was calibrated using offline simulation and engineering data:

[0236] For rapid perturbation threshold By simulating a scenario where generating units (synchronous machines, wind turbines, etc.) suddenly disconnect from the grid, the peak RoCoF is recorded, and the 95th percentile of RoCoF under such scenarios is taken. For example, simulations show that when 20% of the generating units disconnect from the grid, the RoCoF can reach [value missing]. Therefore Set to 8 Hz / s.

[0237] For slow perturbation threshold By simulating load abrupt changes (such as a 10% step change in rated load), the peak value of RoCoF was recorded, and the 95th percentile of RoCoF under such scenarios was taken. For example, when the load suddenly increases by 10%, the RoCoF is approximately... Therefore Set to 2 Hz / s.

[0238] For short-term disturbance windows The generator disconnection is a "transient strong disturbance" with an extremely short duration (typically 200-500ms), therefore Set to 300 ms.

[0239] For long-term disturbance windows The load abruptly changes to a "continuous weak disturbance," which lasts for a relatively long time (usually exceeding 1 second). Set it to 1 s.

[0240] In the disturbance identification process, real-time monitoring With duration ,from Exceed Start timing. If... and This was determined to be a generator disconnection disturbance (a strong, instantaneous disturbance requiring rapid replenishment of large inertia). If and It was determined to be a load change disturbance (a continuous weak disturbance that requires continuous replenishment of inertia to maintain frequency stability).

[0241] S63. Based on the real-time frequency change rate, disturbance type identifier, and preset weighting coefficients, the virtual inertia of the microgrid is calculated using a weighted evaluation model to obtain the system virtual inertia of the microgrid. The formula for calculating the virtual inertia of the microgrid is as follows:

[0242]

[0243] in, To meet the demand for virtual inertia, The frequency change rate weighting coefficient, For the weighting coefficient of changing acceleration, This indicates the disturbance type impact factor set according to the disturbance type identifier. It represents the acceleration due to frequency change.

[0244] Specifically, the weighted evaluation model integrates three dimensions—frequency change rate amplitude, disturbance type, and frequency change acceleration—to quantify the microgrid's demand intensity for virtual inertia.

[0245] Frequency change rate weighting coefficient The direct contribution of RoCoF to inertia demand can be calibrated through a simulation experiment of "inertia-frequency deviation". In the microgrid simulation model, the virtual inertia J is gradually increased, and the maximum frequency deviation caused by RoCoF under different J values ​​is recorded. ,Establish The linear relationship was fitted using the least squares method. and ,but This ensures that the inertia requirement and RoCoF have a reasonable linear relationship.

[0246] Change acceleration weighting coefficient The RoCoF trend reflects the impact on inertia requirements. An increase in the absolute value of RoCoF indicates faster frequency degradation, requiring more inertia. Calibration. The steps are as follows: . Then

[0247] Disturbance type influence factor Adjust the inertia demand weight according to the type of disturbance. If it is a unit disconnection disturbance, it can be set to... This indicates that a strong instantaneous disturbance requires more inertia to quickly suppress the frequency drop; if it is a sudden load change disturbance, it can be set to... This indicates that a sustained weak disturbance requires more sustained inertia support.

[0248] In this implementation, by accurately identifying the type of microgrid disturbance and dynamically assessing inertial demand, the calculation of virtual inertia demand is made to better reflect the actual disturbance scenario, thereby improving the accuracy of inertial demand assessment.

[0249] The aforementioned multi-energy collaborative smart microgrid control method accurately assesses the system's inertia demand by real-time acquisition of microgrid frequency signals and calculation of their rate of change. It then dynamically calculates the maximum virtual inertia each unit can provide, considering the physical characteristics and operational constraints of energy units such as wind turbines, photovoltaics, and energy storage. Using grey relational analysis, it dynamically matches system demand with unit capabilities to generate an optimal allocation scheme. This drives the dedicated control models of each unit to generate differentiated power regulation commands. Finally, power support is collaboratively executed through power electronic converters. This method achieves precise matching and efficient collaboration of multi-energy virtual inertia capabilities, significantly enhancing the frequency stability of high-proportion renewable energy grids and effectively avoiding problems such as resource allocation mismatch and poor control adaptability in traditional methods. It also ensures the dynamic safety of the system under complex operating conditions.

[0250] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0251] Based on the same inventive concept, this application also provides an apparatus for implementing the multi-energy coordinated smart microgrid control method described above. The solution provided by this apparatus is similar to the implementation scheme described in the above method; therefore, the specific limitations of one or more multi-energy coordinated smart microgrid control apparatus embodiments provided below can be found in the limitations of the multi-energy coordinated smart microgrid control method described above, and will not be repeated here.

[0252] In one exemplary embodiment, such as Figure 3 As shown, a multi-energy coordinated smart microgrid control device 30 is provided to implement the methods in the above-described method embodiments. The device includes:

[0253] The dynamic inertial demand assessment module 31 is used to collect the frequency signal of the microgrid system through the synchronous phasor measurement unit deployed at the key nodes of the microgrid; based on the frequency change rate of the frequency signal, the virtual inertia demand of the microgrid system is calculated through inertial demand assessment.

[0254] The energy unit inertia capability assessment module 32 is used to calculate the maximum virtual inertia that each energy unit can provide based on the physical model and operating constraints of each energy unit in the microgrid system.

[0255] The virtual inertia allocation optimization module 33 is used to calculate the actual allocated virtual inertia of each energy unit based on the required virtual inertia and the maximum available virtual inertia using grey relational analysis.

[0256] The additional power command generation module 34 is used to generate additional power commands for each energy unit based on the actual allocated virtual inertia and the corresponding operation control model of each energy unit.

[0257] The collaborative power regulation execution module 35 is used to superimpose the additional power commands corresponding to each energy unit onto the local control loop of the microgrid system, drive the power electronic converter to perform power regulation, and perform virtual inertial control of multi-energy collaboration.

[0258] Embodiments of this application also provide a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the aforementioned method embodiments.

[0259] Embodiments of this application also provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the above-described method embodiments.

[0260] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The components described as separate parts may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this disclosure according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0261] The above-described embodiments are merely illustrative of several implementation methods of the embodiments of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the embodiments of this application, and these modifications and improvements all fall within the protection scope of the embodiments of this application.

Claims

1. A multi-energy coordinated intelligent microgrid control method, characterized in that, The method includes: S1. Frequency signals of the microgrid system are acquired by synchronous phasor measurement units deployed at key nodes of the microgrid; based on the frequency change rate of the frequency signals, the required virtual inertia of the microgrid system is calculated through inertial demand assessment. S2. Based on the physical model and operating constraints of each energy unit in the microgrid system, calculate the maximum virtual inertia that each energy unit can provide; S3. Based on the required virtual inertia and the maximum available virtual inertia, the actual allocated virtual inertia of each energy unit is calculated using grey relational analysis. S4. Based on the actual allocated virtual inertia, generate additional power commands corresponding to each energy unit through the operation control model corresponding to each energy unit; S5. The additional power command corresponding to each energy unit is superimposed on the local control loop of the microgrid system to drive the power electronic converter to perform power regulation in order to perform virtual inertial control of multi-energy coordination. The calculation of the virtual inertia demand of the microgrid system based on the frequency change rate of the frequency signal through inertial demand assessment includes: S61. Based on the frequency signal, the frequency change rate is calculated using the center difference method to obtain the real-time frequency change rate; the formula for calculating the real-time frequency change rate is: ； Wherein, RoCoF is the real-time frequency change rate. This represents the frequency value at time t+Δt. This represents the frequency value at time t-Δt, where Δt is the sampling interval; S62. Based on the real-time frequency change rate and the corresponding duration, identify the disturbance type identifier; Among them, when And duration When, it is determined to be a unit disconnection disturbance; when and When this occurs, it is determined to be a sudden load disturbance; among which, To quickly set the perturbation threshold, The slow disturbance threshold, For short-term disturbance windows, For long-term disturbance windows; S63. Based on the real-time frequency change rate, the disturbance type identifier, and the preset weighting coefficient, the required virtual inertia is calculated using a weighted evaluation model to obtain the system's required virtual inertia for the microgrid; the calculation formula for the required virtual inertia is: ； in, For the required virtual inertia, The frequency change rate weighting coefficient, For the weighting coefficient of changing acceleration, This indicates the disturbance type influence factor set according to the disturbance type identifier. It represents the acceleration due to frequency change.

2. The method according to claim 1, characterized in that, The energy units include wind turbines, photovoltaic systems, and energy storage systems; the calculation of the maximum virtual inertia that each energy unit can provide, based on the physical model and operating constraints of each energy unit in the microgrid system, includes: S11. Based on the rotor speed and output power of the wind turbine, the currently releaseable rotational kinetic energy is calculated using the wind turbine dynamics model to obtain the current releaseable rotational kinetic energy of the wind turbine. S12. Based on the current releaseable rotational kinetic energy of the wind turbine and the preset maximum allowable rotational speed deviation, calculate the maximum virtual inertia that the wind turbine can provide using the kinetic energy-inertia conversion formula; the kinetic energy-inertia conversion formula is: ； in, This represents the maximum virtual inertia that the wind turbine can provide. The current releaseable rotational kinetic energy of the wind turbine. The current rotor angular velocity of the wind turbine is... The preset maximum allowable speed deviation; S13. Based on the illuminance, operating point voltage, and current data of the photovoltaic system, calculate the maximum available power of the photovoltaic system using the maximum power point tracking algorithm; subtract the current output power of the photovoltaic system from the maximum available power to obtain the power margin of the photovoltaic system. S14. Based on the power margin of the photovoltaic system and the preset maximum allowable frequency change rate of the microgrid system, calculate the maximum virtual inertia that the photovoltaic system can provide using a power margin-inertia mapping model; the power margin-inertia mapping model is as follows: ； in, This represents the maximum virtual inertia that the photovoltaic system can provide. The power margin, The angular velocity corresponding to the standard frequency of the microgrid system. This is the preset maximum allowable rate of frequency change; S15. Based on the state of charge, temperature and current charging and discharging current of the energy storage system, calculate the current available charging and discharging power limit of the energy storage system, and subtract the current actual power of the energy storage system from the current available charging and discharging power limit to obtain the charging and discharging power margin of the energy storage system. S16. Based on the charge / discharge power margin of the energy storage system and the preset maximum allowable frequency change rate of the microgrid system, calculate the maximum virtual inertia that the energy storage system can provide using a power-inertia linear conversion model; the power-inertia linear conversion model is as follows: ； in, This represents the maximum virtual inertia that the energy storage system can provide. The charging power margin of the energy storage system. This represents the discharge power margin of the energy storage system.

3. The method according to claim 1, characterized in that, The calculation of the actual allocated virtual inertia for each energy unit based on the required virtual inertia and the maximum available virtual inertia, using grey relational analysis, includes: S21. Based on the maximum available virtual inertia and response characteristic indicators of each energy unit, construct a comparison sequence for each energy unit; the comparison sequence is as follows: ;in, This represents the comparison sequence for the i-th energy unit; This represents the response characteristic index of the i-th energy unit; This represents the maximum virtual inertia that the i-th energy unit can provide; S22. Calculate the correlation coefficient between the comparison sequence of each energy unit and the preset ideal sequence using the grey relational coefficient formula, and use this coefficient as the grey relational coefficient of each energy unit; the grey relational coefficient formula is: ； in, The correlation coefficient of the i-th energy unit on the k-th dimension feature of the comparison sequence is represented by k=1 or 2. When k=1, it represents the response characteristic dimension, and when k=2, it represents the maximum available inertia dimension. This represents the characteristic value of the k-th dimension of the ideal sequence. The characteristic value of the k-th dimension of the comparison sequence of the i-th energy unit; , ρ is the preset resolution coefficient; S23. Based on the grey relational coefficients of each energy unit, the grey relational degree of each energy unit is calculated using an arithmetic average algorithm; the formula for calculating the grey relational degree is: ； in, The gray relational degree of the i-th energy unit is given by N, where N is the number of feature indicators of the comparison sequence. S24. Based on the grey relational degree and the maximum available virtual inertia of each energy unit, calculate the allocation coefficient of each energy unit using a proportional allocation formula; the proportional allocation formula is: ； in, This represents the allocation coefficient for the i-th energy unit; S25. Based on the allocation coefficient and the required virtual inertia, calculate the actual allocated virtual inertia of each energy unit using a linear allocation formula; the linear allocation formula is: ； in, The actual allocation of virtual inertia to the i-th energy unit, The virtual inertia is the requirement.

4. The method according to claim 2, characterized in that, The steps for generating the additional power command corresponding to the wind turbine include: S31. Based on the actual allocated virtual inertia of the wind turbine, the basic additional power command of the wind turbine is generated by using the synchronous generator rotor motion equation model and combining real-time frequency change rate and frequency deviation data. S32. Based on the basic additional power command of the wind turbine and the rotor speed change trend, the basic additional power command is dynamically corrected when a speed recovery requirement is detected through the speed recovery limiting mechanism to obtain the corrected additional power command of the wind turbine. S33. By superimposing the corrected additional power command onto the original power reference value through the power control loop of the wind turbine rotor-side converter, the final active power regulation command of the wind turbine is obtained.

5. The method according to claim 2, characterized in that, The steps for generating the additional power command corresponding to the photovoltaic system include: S41. Based on the actual allocated virtual inertia of the photovoltaic system, generate the basic additional power command of the photovoltaic system through the virtual inertia-power conversion model; S42. Based on historical frequency disturbance statistics and future light intensity predictions, calculate the current power reserve margin required by the photovoltaic system using a power margin adaptive algorithm. S43. Based on the basic additional power command of the photovoltaic system and the power reserve margin, the final active power adjustment command of the photovoltaic system is generated through superposition calculation.

6. The method according to claim 2, characterized in that, The steps for generating the additional power command corresponding to the energy storage system include: S51. The actual virtual inertia of the energy storage system is allocated into fast response inertia and continuous support inertia according to a preset ratio. S52. Based on the fast response inertia, generate additional power commands for the fast response stage through the differential control loop. S53. Based on the continuous support inertia, generate additional power commands for the continuous support stage through a proportional control loop. S54. The additional power command of the rapid response phase and the additional power command of the continuous support phase are superimposed and calculated to obtain the total additional power command of the energy storage system. S55. Based on the total additional power command and real-time state of charge of the energy storage system, the amplitude of the total additional power command is constrained by a dynamic limiting algorithm to obtain the final active power regulation command of the energy storage system.

7. A multi-energy coordinated intelligent microgrid control device, used to implement the method according to any one of claims 1 to 6, characterized in that, The device includes: The dynamic inertial demand assessment module is used to collect the frequency signal of the microgrid system through synchronous phasor measurement units deployed at key nodes of the microgrid; based on the frequency change rate of the frequency signal, the required virtual inertia of the microgrid system is calculated through inertial demand assessment. The energy unit inertia capability assessment module is used to calculate the maximum virtual inertia that each energy unit can provide based on the physical model and operating constraints of each energy unit in the microgrid system. The virtual inertia allocation optimization module is used to calculate the actual allocated virtual inertia of each energy unit based on the required virtual inertia and the maximum available virtual inertia, using grey relational analysis. An additional power command generation module is used to generate additional power commands corresponding to each energy unit based on the actual allocated virtual inertia and through the operation control model corresponding to each energy unit. The collaborative power regulation execution module is used to superimpose the additional power commands corresponding to each energy unit onto the local control loop of the microgrid system, drive the power electronic converter to perform power regulation, and perform virtual inertial control of multi-energy collaboration. The calculation of the virtual inertia demand of the microgrid system based on the frequency change rate of the frequency signal through inertial demand assessment includes: S61. Based on the frequency signal, the frequency change rate is calculated using the center difference method to obtain the real-time frequency change rate; the formula for calculating the real-time frequency change rate is: ； Wherein, RoCoF is the real-time frequency change rate. This represents the frequency value at time t+Δt. This represents the frequency value at time t-Δt, where Δt is the sampling interval; S62. Based on the real-time frequency change rate and the corresponding duration, identify the disturbance type identifier; Among them, when And duration When, it is determined to be a unit disconnection disturbance; when and When this occurs, it is determined to be a sudden load disturbance; among which, To quickly set the perturbation threshold, The slow disturbance threshold, For short-term disturbance windows, For long-term disturbance windows; S63. Based on the real-time frequency change rate, the disturbance type identifier, and the preset weighting coefficient, the required virtual inertia is calculated using a weighted evaluation model to obtain the system's required virtual inertia for the microgrid; the calculation formula for the required virtual inertia is: ； in, For the required virtual inertia, The frequency change rate weighting coefficient, For the weighting coefficient of changing acceleration, This indicates the disturbance type influence factor set according to the disturbance type identifier. It represents the acceleration due to frequency change.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 6.

Images