A self-synchronous voltage source energy storage controller parameter optimization setting method and system

Abstract

The application provides a self-synchronous voltage source energy storage controller parameter optimization setting method and system, and belongs to the technical field of smart grids, and comprises the following steps: continuously monitoring the active power fluctuation amplitude of an energy storage grid-connected point; when the active power fluctuation amplitude exceeds a trigger threshold, recording the current time as a trigger time, and intercepting the grid-connected point frequency change curve within a preset time length from the trigger time as a starting point; calculating the system total inertia time constant identification value under the current power grid operation state; calculating the system inertia demand value required for maintaining system frequency safety, and determining the system inertia shortage amount; inputting the system inertia shortage amount into an integral feedback controller, outputting the obtained energy storage virtual inertia time correction amount, and limiting the virtual inertia time correction amount with the rated power capacity of the energy storage converter as an amplitude limiting condition; synchronously updating the droop coefficient to obtain an updated control parameter combination. The application solves the problem that the energy storage controller parameters cannot be adaptively matched according to the power grid strength change.

Description

Technical Field

[0001] This invention relates to the field of smart grid technology, specifically to a method and system for optimizing the parameters of a self-synchronizing voltage source energy storage controller. Background Technology

[0002] With a high proportion of renewable energy integrated into the grid, the power grid exhibits characteristics of weak inertia and weak damping. Self-synchronizing voltage source energy storage controllers need to maintain frequency stability under different grid strengths. Grid strength is typically measured by the short-circuit ratio, and changes in the short-circuit ratio directly affect the system's damping and inertia response characteristics, requiring the energy storage controller to possess corresponding adaptive adjustment capabilities.

[0003] In existing technologies, when facing changes in grid strength, a parameter group switching strategy based on lookup tables is typically adopted. This involves pre-setting multiple sets of virtual inertia and damping parameters for different short-circuit ratio ranges, and passively switching the corresponding parameter group based on the current short-circuit ratio range during operation. This method has the following drawbacks: the parameter groups are preset offline and limited in number; the switching method is passive, making it impossible to achieve real-time continuous matching between parameters and grid strength; when the short-circuit ratio changes, fixed parameters can easily lead to increased transient frequency overshoot and prolonged frequency recovery time; under weak grid conditions, mismatched combinations of inertia and damping parameters may cause power oscillations or even instability.

[0004] In summary, existing lookup table switching strategies cannot achieve real-time adaptive matching between energy storage controller parameters and grid strength, and are difficult to balance frequency response speed and system stability in scenarios with dynamically changing short-circuit ratios. Summary of the Invention

[0005] This invention provides a method and system for optimizing the parameters of a self-synchronizing voltage source energy storage controller, which solves the problem in the prior art that the parameters of the energy storage controller cannot be adaptively matched according to changes in grid strength.

[0006] In view of the above problems, the present invention provides a method and system for optimizing the parameters of a self-synchronizing voltage source energy storage controller.

[0007] In a first aspect, the present invention provides a method for optimizing and tuning the parameters of a self-synchronizing voltage source energy storage controller, comprising: S1: Continuously monitor the active power fluctuation amplitude of the energy storage grid connection point. When the active power fluctuation amplitude exceeds the trigger threshold, record the current time as the trigger time, and extract the grid connection point frequency change curve within a preset time period starting from the trigger time. S2: Based on the peak value of the frequency change rate of the frequency change curve at the grid connection point within the first power frequency cycle after the triggering time, and combined with the active power fluctuation amplitude, calculate the identification value of the total system inertia time constant under the current power grid operating state; S3: Based on the current rated frequency of the power grid and the maximum allowable rate of frequency change constraint, calculate the system inertia requirement value required to maintain system frequency security, and determine the system inertia shortage by combining the system total inertia time constant identification value; S4: Input the system inertia shortage into the integral feedback controller, output the virtual inertia time correction of the energy storage, and limit the virtual inertia time correction of the energy storage converter with the rated power capacity of the energy storage converter as the limiting condition; S5: Update the virtual inertia time constant of the energy storage according to the virtual inertia time correction after the limit, and update the droop coefficient synchronously based on the virtual inertia time constant before and after the update to obtain the updated control parameter combination.

[0008] Secondly, the present invention provides a parameter optimization and tuning system for a self-synchronizing voltage source energy storage controller, comprising: The event detection and data capture module is used to continuously monitor the active power fluctuation amplitude of the energy storage grid connection point. When the active power fluctuation amplitude exceeds the trigger threshold, the current time is recorded as the trigger time, and the frequency change curve of the grid connection point within a preset time period is captured from the trigger time. The total inertia online identification module is used to calculate the system total inertia time constant identification value under the current power grid operating state based on the peak value of the frequency change rate of the frequency change curve at the grid connection point within the first power frequency cycle after the triggering time, combined with the active power fluctuation amplitude. The inertia shortage assessment module is used to calculate the system inertia requirement required to maintain system frequency security based on the current grid rated frequency and the maximum allowable rate of frequency change constraint, and to determine the system inertia shortage by combining the system total inertia time constant identification value. The correction calculation and limiting module is used to input the system inertia shortage into the integral feedback controller, output the virtual inertia time correction of the energy storage, and limit the virtual inertia time correction of the energy storage converter with the rated power capacity of the energy storage converter as the limiting condition. The control parameter update module is used to update the virtual inertia time constant of the energy storage according to the virtual inertia time correction after the limit, and to update the droop coefficient synchronously based on the virtual inertia time constant before and after the update, so as to obtain the updated control parameter combination.

[0009] One or more technical solutions provided in this invention have at least the following technical effects or advantages: The technical solution of this invention firstly monitors the fluctuation amplitude of active power at the energy storage grid connection point, and records the triggering time and extracts the frequency change curve of the grid connection point when the fluctuation exceeds the adaptive triggering threshold. This achieves reliable detection of disturbance events and accurate capture of frequency dynamic data, providing complete and effective input data for subsequent online identification of system inertia.

[0010] Furthermore, based on the extracted frequency change curve, the frequency sampling sequence within the first power frequency cycle after triggering is extracted, the absolute value sequence of the frequency change rate is calculated, and the arithmetic mean of the first N values ​​is taken as the peak value of the frequency change rate. Combined with the disturbance power amplitude and the rated frequency of the power grid, the identification value of the total inertia time constant of the system is calculated, realizing the online quantitative identification of the total inertia level of the system under the current power grid operation state, providing an accurate numerical basis for subsequent inertia shortage assessment.

[0011] Furthermore, based on the maximum allowable rate of change of frequency limit and the rated frequency of the power grid, the basic inertia requirement is calculated. The system inertia requirement is obtained by amplifying the peak value of the rate of change of frequency. The difference between this value and the system total inertia time constant is used to obtain the initial inertia shortage. The system inertia shortage is then obtained by amplifying the value with a voltage correction factor and low-pass filtering. This achieves accurate quantification and smooth output of the current power grid inertia deficiency, providing a stable and reliable input for the integral feedback controller.

[0012] Furthermore, the system inertia shortage is input into the integral feedback controller, which consists of an integral coefficient correction unit, an integral calculation unit, a limiting unit, and an adaptive coefficient update unit. Combined with the available power margin of the energy storage converter, the integral coefficient is corrected, and the virtual inertia time correction is output and limited according to the upper and lower limits of the output. This realizes the adaptive calculation and safe limiting of the virtual inertia time correction of the energy storage, ensuring that the correction is within the power range that the energy storage converter can withstand, and providing a reliable adjustment amount for parameter updates.

[0013] Finally, the virtual inertia time correction after limiting is summed with the current virtual inertia time constant and limited to obtain the updated virtual inertia time constant. Based on the ratio of the virtual inertia time constant before and after the update and the per-unit value of the voltage amplitude at the grid connection point, the voltage support bias factor is calculated, and the droop coefficient is adjusted synchronously. Finally, the updated control parameter combination is obtained and the current parameters in the control loop are replaced. This achieves coordinated and adaptive updating of virtual inertia parameters and droop coefficient, which improves the system's inertia support capability while taking into account the reactive power voltage support requirements.

[0014] In summary, the technical solution of this invention can adjust the energy storage control parameters in real time under the conditions of power grid intensity changes and voltage drops, which not only makes up for the insufficient system inertia, but also enhances the reactive voltage support capability under weak power grid conditions, effectively improving the frequency stability and operational adaptability of the energy storage converter under weak power grid conditions. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating a method for optimizing and tuning parameters of a self-synchronizing voltage source energy storage controller provided by the present invention.

[0016] Figure 2This is a schematic diagram illustrating the calculation process of system inertia shortage in a parameter optimization and tuning method for a self-synchronizing voltage source energy storage controller provided by the present invention.

[0017] Figure 3 This is a schematic diagram of the structure of a self-synchronizing voltage source energy storage controller parameter optimization and tuning system provided by the present invention.

[0018] In the attached diagram, the labels representing each component are as follows: Event detection and data capture module 11, total inertia online identification module 12, inertia shortage assessment module 13, correction calculation and limiting module 14, control parameter update module 15. Detailed Implementation

[0019] This invention provides a method and system for optimizing the parameters of a self-synchronizing voltage source energy storage controller, which solves the problem in the prior art that the parameters of the energy storage controller cannot be adaptively matched according to changes in grid strength.

[0020] It should be noted that the terms "comprising" and "having" are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to these processes, methods, products, or devices.

[0021] Example 1, as Figure 1 As shown, the present invention provides a method for optimizing and tuning the parameters of a self-synchronizing voltage source energy storage controller, the method comprising: S1: Continuously monitor the active power fluctuation amplitude of the energy storage grid connection point. When the active power fluctuation amplitude exceeds the trigger threshold, record the current time as the trigger time, and extract the grid connection point frequency change curve within a preset time period starting from the trigger time.

[0022] This step continuously monitors the fluctuation range of active power at the energy storage grid connection point, and automatically records the trigger time and extracts the frequency change curve of the grid connection point when the fluctuation exceeds the limit.

[0023] Step S1 in the method provided by this invention includes: S1.1: Extract the fundamental components of the instantaneous values ​​of the three-phase voltage and three-phase current at the energy storage grid connection point, obtain the fundamental value of active power, and use a sliding time window to calculate the root mean square value of the active power fluctuation amplitude as the real-time active power fluctuation amplitude. S1.2: Obtain the per-unit value of the voltage amplitude at the grid connection point, multiply the reference trigger threshold by the per-unit value of the voltage amplitude, and obtain the trigger threshold that is adaptively adjusted according to the voltage drop depth; S1.3: When the real-time active power fluctuation amplitude exceeds the adaptive adjustment trigger threshold, the current time is recorded as the trigger time, and the real-time active power fluctuation amplitude is stored as the disturbance power amplitude. S1.4: Starting from the triggering time, extract the grid connection point frequency change curve with a length of two power frequency cycles, wherein the grid connection point frequency change curve is composed of the instantaneous frequency values ​​output by the phase-locked loop arranged continuously according to the sampling points.

[0024] Specifically, at the grid connection point of the energy storage system, three-phase voltage transformers and three-phase current transformers are installed. The primary side of the voltage transformer is connected in parallel to the three-phase bus at the grid connection point, proportionally converting high voltage into low voltage signals. The primary side of the current transformer is connected in series to the three-phase line at the grid connection point, proportionally converting large current into small current signals. The secondary side output signals of the voltage and current transformers are input to an analog-to-digital converter (ADC), which synchronously samples the analog signals at a fixed sampling frequency, converting the continuously changing voltage and current waveforms into discrete digital sequences. Each sampling simultaneously acquires the instantaneous values ​​of the three-phase voltage and three-phase current, and each instantaneous value is the actual instantaneous voltage or current value corresponding to that sampling moment.

[0025] The sampling frequency is usually much higher than the fundamental frequency of the power grid. For example, when the fundamental frequency of the power grid is 50Hz, the sampling frequency can be 10kHz, that is, 200 sampling points are collected in each power frequency cycle to ensure accurate restoration of voltage and current waveforms.

[0026] For example, the rated line voltage of a certain energy storage system's grid connection point is 35kV, and the rated current is 1000A. The voltage transformer ratio is 35000:100, meaning 35kV on the primary side corresponds to 100V on the secondary side; the current transformer ratio is 1000:5, meaning 1000A on the primary side corresponds to 5A on the secondary side. At a certain sampling moment, the instantaneous values ​​of the three-phase voltage on the secondary side collected by the analog-to-digital converter are: Phase A: 80.0V, Phase B: -41.5V, Phase C: -38.5V. Based on the voltage transformer ratios and converted to the primary side, the instantaneous values ​​of the three-phase voltages are: Phase A: 80.0 × 350 = 28000V = 28.0kV, Phase B: -41.5 × 350 = -14525V ≈ -14.5kV, Phase C: -38... 0.5 × 350 = -13475V ≈ -13.5kV. At the same sampling time, the instantaneous values ​​of the three-phase current on the secondary side are: Phase A: 4.0A, Phase B: -2.1A, Phase C: -1.9A. According to the current transformer ratio, the instantaneous values ​​of the three-phase current on the primary side are: Phase A: 4.0 × 200 = 800A, Phase B: -2.1 × 200 = -420A, Phase C: -1.9 × 200 = -380A.

[0027] In this step, step S1.1 first performs Park transformation on the instantaneous values ​​of the three-phase voltage and the three-phase current at the energy storage grid connection point. The basic principle of Park transformation is to project the sinusoidal electrical quantities in the stationary three-phase coordinate system into two orthogonal DC components through a coordinate system that rotates synchronously with the grid voltage.

[0028] Specifically, the Parker transform utilizes the real-time phase angle of the grid voltage output from the phase-locked loop to construct a rotating dq coordinate system. Taking the direction of the grid voltage vector as the d-axis direction, the three-phase instantaneous values ​​are projected according to this phase angle: the component in the same direction as the voltage vector is output as the d-axis component, representing the active component; the component perpendicular to the voltage vector is output as the q-axis component, representing the reactive component. For an ideal three-phase symmetrical fundamental voltage, the transformed d-axis component is a constant DC quantity in steady state, and the q-axis component is approximately zero.

[0029] Perform a Parker transformation on the three-phase voltages to obtain the d-axis voltage components and q-axis voltage components; perform a Parker transformation on the three-phase currents to obtain the d-axis current components and q-axis current components.

[0030] For example, the real-time phase angle of the grid voltage output by the phase-locked loop (PLL) is 0 degrees. After Parker transformation, the instantaneous three-phase voltage values ​​of 28.0kV, -14.5kV, and -13.5kV have a d-axis voltage component of approximately 28.5kV and a q-axis voltage component of approximately 0kV. Similarly, after Parker transformation, the instantaneous three-phase current values ​​of 800A, -420A, and -380A have a d-axis current component of approximately 1.2kA and a q-axis current component of approximately 0kA. The 28.5kV d-axis voltage component is a DC quantity, representing the amplitude of the fundamental active component of the voltage; the near-zero q-axis voltage component indicates good PLL tracking.

[0031] Furthermore, when there is voltage asymmetry or disturbance at the grid connection point, the three-phase voltage will contain a negative-sequence component. The negative-sequence component rotates in the opposite direction to the positive-sequence component. After the Parker transform, it manifests as an AC ripple at twice the fundamental frequency superimposed on the DC component in the d-axis and q-axis components. For example, if the grid fundamental frequency is 50Hz, the ripple frequency introduced by the negative-sequence component is 100Hz. This ripple interferes with the accurate calculation of the fundamental amplitude and power.

[0032] To accurately extract the fundamental frequency component, a first-order low-pass filter with a cutoff frequency twice the fundamental frequency is used to filter the obtained d-axis voltage component, q-axis voltage component, d-axis current component, and q-axis current component. The cutoff frequency is set to twice the fundamental frequency of the power grid; for example, when the fundamental frequency is 50Hz, the cutoff frequency is 100Hz. The transfer function of the first-order low-pass filter is 1 / (τs+1), where τ is the time constant, τ=1 / (2πfc), and fc is the cutoff frequency. In the digital implementation, backward Euler discretization is used, and the recursive formula is: y(k)=α·x(k)+(1-α)·y(k-1), where α=2πfc·Ts / (1+2πfc·Ts), Ts is the sampling period, x(k) is the current sampled input value, y(k) is the current filtered output value, and y(k-1) is the filtered output value at the previous sampling time. A first-order low-pass filter has an attenuation characteristic of 20dB per decade for 100Hz AC signals. After filtering, the second harmonic ripple is significantly suppressed, while the DC signal passes through without attenuation.

[0033] For example, the fundamental frequency of the power grid is 50Hz, and the cutoff frequency of the first-order low-pass filter is set to 100Hz. Assume that the d-axis voltage component has a 100Hz ripple with a peak-to-peak value of 1.0kV superimposed on it, with an original value of 28.5kV ± 0.5kV fluctuation. After the first-order low-pass filter, the 100Hz ripple is attenuated by approximately 3dB, and its amplitude is reduced to approximately 0.707 times. After filtering, the d-axis voltage component is approximately 28.5kV, with a ripple peak fluctuation of ±0.35kV, meaning the peak-to-peak ripple value is 0.7kV. The DC component remains unchanged at 28.5kV. After filtering, the q-axis voltage component is approximately 0kV, the d-axis current component is approximately 1.2kA, and the q-axis current component is approximately 0kA. The second harmonic ripple in each component is effectively filtered out.

[0034] Furthermore, the fundamental value of active power is calculated using the filtered d-axis and q-axis voltage and current components. According to instantaneous power theory, under the constant amplitude Parker transform, the fundamental value of active power is equal to three-half of the product of the d-axis voltage filter value and the d-axis current filter value, plus the product of the q-axis voltage filter value and the q-axis current filter value.

[0035] A sliding time window with a length of several sampling points is used to calculate the root mean square (RMS) fluctuation amplitude of the fundamental active power value in real time. The sliding step size of the sliding time window is one sampling point; that is, as soon as a new sampling point is acquired, the window moves forward by one sampling point, discarding the oldest sampling point within the window, thus achieving point-by-point updating. Within the sliding time window, the deviation of the fundamental active power value of each sampling point from the average value within the window is calculated. The arithmetic mean of the squares of the deviations is then taken, and the square root of this average value is the real-time active power fluctuation amplitude.

[0036] For example, after filtering, the d-axis voltage is 28.5kV, the q-axis voltage is 0kV, the d-axis current is 1.2kA, and the q-axis current is 0kA. Before filtering, each of these components also has second-harmonic ripple of varying amplitudes superimposed on it. After processing by a first-order low-pass filter with the same cutoff frequency of 100Hz, the peak ripple value of each component is attenuated to approximately 0.707 times its original value, while the DC component remains unchanged. The fundamental active power value is approximately 1.5 × (28500 × 1200 + 0 × 0) ≈ 51.30MW. The sliding time window contains 100 sampling points. The fundamental active power values ​​of the five most recent sampling points within the window are 51.2, 51.5, 50.8, 51.0, and 51.3MW, with an average value of 51.16MW. The squared deviations at each point are 0.0016, 0.1156, 0.1296, 0.0256, and 0.0196MW², respectively. The arithmetic mean of the squared deviation is 0.0584 MW², and the real-time active power fluctuation after taking the square root is approximately 0.24 MW.

[0037] Further, in step S1.2, the per-unit value of the voltage amplitude at the grid connection point is obtained, and the reference trigger threshold is multiplied by the per-unit value of the voltage amplitude to obtain the trigger threshold that is adaptively adjusted according to the voltage drop depth.

[0038] In this step, the step of obtaining the per-unit value of the grid connection point voltage amplitude includes: Input the instantaneous three-phase voltage values ​​at the energy storage grid connection point into the Parker transformation matrix, and output the d-axis voltage components and q-axis voltage components in the rotating coordinate system; The d-axis voltage component and the q-axis voltage component are filtered by a first-order low-pass filter with a cutoff frequency of twice the fundamental frequency, respectively, to obtain the filtered d-axis voltage component and the filtered q-axis voltage component. Calculate the sum of squares of the filtered d-axis voltage component and the filtered q-axis voltage component, and take the square root of the sum of squares to obtain the fundamental amplitude of the grid connection point voltage. A moving average window is constructed with half a power frequency cycle as the window length. The fundamental amplitude of the grid connection point voltage is processed by moving average, and the result after moving average processing is used as the smoothed voltage amplitude. Divide the smoothed voltage amplitude by the AC voltage reference value to obtain the per-unit value of the grid connection point voltage amplitude, wherein the AC voltage reference value is the phase voltage amplitude reference value corresponding to the rated line voltage of the grid connection point.

[0039] Specifically, the sum of the squares of the filtered d-axis voltage component and the filtered q-axis voltage component is first calculated, and the square root of the sum of squares is taken to obtain the fundamental amplitude of the grid connection point voltage.

[0040] The above calculation is based on the definition of voltage vector amplitude in a rotating coordinate system: the voltage vector amplitude is equal to the length of the hypotenuse of the right triangle formed by the d-axis component and the q-axis component. Since the low-pass filter has filtered out the AC ripple introduced by the negative sequence, this fundamental amplitude can accurately reflect the true amplitude of the positive sequence fundamental voltage at the grid connection point.

[0041] For example, if the filtered d-axis voltage is 28.5kV and the q-axis voltage is 0kV, then the fundamental voltage amplitude = ≈28.50kV.

[0042] Secondly, a moving average window is constructed with a window length of half a power frequency cycle to perform moving average processing on the obtained fundamental amplitude values ​​of the grid-connected voltage. Half a power frequency cycle corresponds to 0.01 seconds in a 50Hz power grid, and the window contains the fundamental amplitude values ​​of all sampling points within that time period. The arithmetic mean of the fundamental amplitude values ​​of each sampling point within the window is taken as the smoothed voltage amplitude at the current moment. As time progresses, the window slides forward, continuously updating the smoothing results. The sliding step size of the moving average window is 1 sampling point; each time a new fundamental amplitude sampling point is obtained, the window moves forward by one sampling point.

[0043] The purpose of the above-mentioned moving average processing is to further suppress random fluctuations and residual ripple interference in the fundamental amplitude, making the output voltage amplitude change smoother and avoiding the trigger threshold from malfunctioning due to instantaneous spikes.

[0044] For example, with a sampling frequency of 10kHz, half a power frequency cycle (0.01 seconds) contains 100 sampling points. The fundamental amplitudes of the most recent 10 sampling points within the window are 28.50, 28.51, 28.49, 28.50, 28.52, 28.48, 28.50, 28.51, 28.49, and 28.50 kV, respectively. The arithmetic mean is 28.50 kV, meaning the smoothed voltage amplitude is 28.50 kV.

[0045] Next, the smoothed voltage amplitude is divided by the AC voltage reference value to obtain the per-unit value of the grid connection point voltage amplitude. The per-unit value is a dimensionless numerical value that represents the relative magnitude of electrical quantities in a power system. 1.0 indicates that the voltage is equal to the rated value, and less than 1.0 indicates a voltage drop.

[0046] The AC voltage reference value is the phase voltage amplitude reference value corresponding to the rated line voltage at the grid connection point. Its calculation formula is: √2 divided by √3, then multiplied by the effective value of the rated line voltage at the grid connection point. √2 divided by √3 is approximately equal to 0.8165; this coefficient converts the effective line voltage value into the phase voltage amplitude reference value.

[0047] The effective value of the rated line voltage at the grid connection point is obtained from the design specifications of the energy storage system's grid connection interface. In this step, 35kV is used, which is the rated voltage level of the power grid bus directly connected to the energy storage system, as the basis for calculating the AC voltage reference value. Substituting into the formula, the AC voltage reference value is approximately 28578 volts.

[0048] For example, if the effective value of the rated line voltage at the grid connection point is 35kV, and the AC voltage reference value is 0.8165×35000≈28.58kV, then the per-unit value of the voltage amplitude is 28.50 / 28.58≈0.997.

[0049] Furthermore, the preset reference trigger threshold is multiplied by the per-unit value of the grid connection point voltage amplitude to obtain the trigger threshold that adaptively adjusts with the voltage drop depth. The reference trigger threshold is set according to the percentage of the rated power of the energy storage system.

[0050] For example, if the energy storage system has a rated power of 50MW, the baseline trigger threshold can be set to 10% of the rated power, i.e., 5.0MW. This percentage can be adjusted according to the short-circuit capacity at the grid connection point, voltage level, and grid operation procedures, with a typical setting range of 5% to 20% of the rated power.

[0051] Further, step S1.3 compares the real-time active power fluctuation amplitude with the adaptive trigger threshold in real time. When the fluctuation amplitude exceeds the trigger threshold, the current time is recorded as the trigger time, and the current real-time active power fluctuation amplitude is stored as the disturbance power amplitude. When the fluctuation amplitude does not exceed the trigger threshold, the monitoring continues cyclically. When the per-unit voltage amplitude at the grid connection point is lower than 0.1, the trigger threshold is set to the lower limit, which is 10% of the base trigger threshold. As the per-unit voltage amplitude further drops from 0.1 to 0, the trigger threshold remains unchanged at this lower limit and does not continue to decrease with the per-unit voltage amplitude, thus avoiding frequent false triggering due to an excessively low adaptive trigger threshold under deep voltage drops or zero voltage conditions.

[0052] For example, the baseline trigger threshold is set to 5.0 MW. Scenario 1: Voltage is normal, per-unit value is 0.997, trigger threshold = 5.0 × 0.997 = 4.985 MW. The current fluctuation amplitude is 0.24 MW < 4.985 MW, so no trigger is given. Scenario 2: Voltage drops to 0.70 per-unit value. Trigger threshold = 5.0 × 0.70 = 3.5 MW. At this time, the system experiences power oscillation due to a fault, with a fluctuation amplitude of 4.8 MW > 3.5 MW, triggering the system. The trigger time is recorded at the 12500th sampling point, i.e., at 1.25 seconds, and the disturbance power amplitude is stored as 4.8 MW.

[0053] Finally, step S1.4 uses the recorded trigger time as the starting point and extracts a grid connection point frequency change curve with a length of two power frequency cycles. Two power frequency cycles correspond to 0.04 seconds in a 50Hz power grid. The extraction interval is [trigger time, trigger time + 0.04 seconds]. The trigger time is the sampling point corresponding to the moment when the active power fluctuation amplitude exceeds the adaptive trigger threshold. The instantaneous frequency value of this sampling point is used as the first sampling point of the extracted curve. Data before the trigger time is not included in the extraction range. This frequency change curve is composed of the instantaneous frequency values ​​output by the phase-locked loop arranged continuously according to the sampling interval. The extracted curve segment contains the complete frequency dynamic characteristics after the trigger time for use in subsequent steps.

[0054] For example, taking scenario two above, the trigger time is 1.25 seconds, the sampling frequency is 10kHz, and two power frequency cycles contain 400 sampling points. The instantaneous frequency sequence of the phase-locked loop (PLL) from 1.25 seconds to 1.29 seconds is captured, totaling 400 points. The sampling point frequency corresponding to the trigger time is 50.00Hz. The frequency then continuously decreases, reaching 49.92Hz approximately 0.008 seconds after triggering, corresponding to the point of maximum frequency change. During this period, the frequency drops from 50.00Hz to 49.92Hz, a change of -0.08Hz, with an average frequency change rate of -0.08Hz / 0.008s = -10Hz / s. The frequency then continues to decrease, reaching its lowest point of 49.90Hz approximately 0.02 seconds after triggering, before starting to rise again, reaching 49.98Hz at the end of the capture window. This dynamic trajectory then proceeds to the next steps.

[0055] It should be noted that the above values ​​are for illustrative purposes only and do not constitute a limitation on the present invention.

[0056] In summary, this step achieves reliable detection of disturbance events and accurate capture of frequency dynamic data, providing complete and effective input data for subsequent online identification of system inertia.

[0057] S2: Based on the peak value of the frequency change rate of the frequency change curve at the grid connection point within the first power frequency cycle after the triggering time, and combined with the active power fluctuation amplitude, calculate the identification value of the total system inertia time constant under the current power grid operating state.

[0058] This step involves extracting frequency sampling data within the first power frequency cycle after triggering and calculating the peak value of the frequency change rate, then combining this with the disturbance power amplitude to calculate the identification value of the system's total inertial time constant.

[0059] Step S2 in the method provided by this invention includes: S2.1: Extract the frequency sampling point sequence within the first power frequency cycle after the triggering time from the frequency change curve of the grid connection point, and arrange them in order of sampling time; S2.2: Calculate the absolute value of the difference between instantaneous frequency values ​​between adjacent sampling points in the frequency sampling point sequence to obtain a sequence of absolute values ​​of frequency change rate; S2.3: Sort each value in the absolute value sequence of frequency change rate in descending order, and take the arithmetic mean of the first N values ​​after sorting as the peak value of frequency change rate, where N is determined according to the total number of sampling points in the first power frequency cycle; S2.4: Obtain the disturbance power amplitude, and combine it with the grid rated frequency and the peak value of the frequency change rate to obtain the system total inertia time constant identification value in terms of time.

[0060] Specifically, firstly, step S2.1 extracts the frequency sampling point sequence within the first power frequency cycle after the triggering time from the frequency change curve of the grid connection point intercepted in step S1, and arranges them in the order of sampling time.

[0061] The start and end times of the first power frequency cycle are defined as follows: the start point is the sampling point corresponding to the trigger moment, and the end point is the sampling point corresponding to the next power frequency cycle after the trigger moment. In a 50Hz power grid, one power frequency cycle is 0.02 seconds. When the sampling frequency is 10kHz, one power frequency cycle contains 200 sampling points, therefore the extracted sequence length is 200 frequency sampling points.

[0062] For example, in step S1, the trigger time is 1.25 seconds, and the frequency change curve range is from 1.25 seconds to 1.29 seconds. A frequency sampling point sequence from 1.25 seconds to 1.27 seconds is extracted from this curve, totaling 200 sampling points, arranged in chronological order. f(1) is the frequency sampling value of 50.00Hz corresponding to the trigger time, and f(200) is the frequency sampling value corresponding to 0.02 seconds after the trigger.

[0063] Further, in step S2.2, the absolute value of the difference in instantaneous frequency values ​​between adjacent sampling points is calculated from the frequency sampling point sequence extracted in S2.1 to obtain a sequence of absolute values ​​of the frequency change rate. For two adjacent sampling points f(k) and f(k+1), the difference in instantaneous frequency values ​​is f(k+1) - f(k). This difference divided by the sampling period Ts is the average frequency change rate between adjacent sampling points, in Hz / s. The sampling period Ts is the reciprocal of the sampling frequency; when the sampling frequency is 10kHz, Ts is 0.0001 seconds.

[0064] The absolute value of the rate of change of frequency is calculated using the following formula: R(k) = |f(k+1) - f(k)| / Ts, where f(k) is the instantaneous frequency value at the k-th sampling point, f(k+1) is the instantaneous frequency value at the (k+1)-th sampling point, Ts is the sampling period, and R(k) is the absolute value of the rate of change of frequency at the k-th sampling point, in Hz / s. The sequence of 200 sampling points contains 199 pairs of adjacent sampling points, corresponding to 199 absolute values ​​of the rate of change of frequency, forming a sequence of absolute values ​​of the rate of change of frequency.

[0065] For example, in the sequence extracted in the previous step, f(1) = 50.00Hz, f(2) = 49.99Hz, and Ts = 0.0001s. Then R(1) = |49.99 - 50.00| / 0.0001 = 100Hz / s. Continuing to calculate subsequent adjacent points, a sequence of absolute values ​​of the rate of change of frequency of 199 values ​​is obtained.

[0066] Next, in step S2.3, the values ​​in the absolute value sequence of frequency change rate obtained in S2.2 are sorted in descending order, and the arithmetic mean of the first N values ​​after sorting is taken as the peak value of frequency change rate.

[0067] N is determined based on the total number of sampling points within the first power frequency cycle. When the total number of sampling points is 200, N is 5% of the total number of sampling points, i.e., N=10. The principle for determining N is: if N is too small, the peak value is easily affected by a single abnormal sampling point; if N is too large, the average value is lowered by subsequent smaller values, failing to reflect the true peak value. The 5% value strikes a balance between suppressing sampling noise and maintaining peak representativeness.

[0068] After sorting in descending order, the first N values ​​are the N largest absolute values ​​of frequency change rates in the sequence. The arithmetic mean of these N values ​​is the peak frequency change rate, denoted as Rp, with units of Hz / s.

[0069] For example, among the 199 absolute values ​​of frequency change rate obtained in the previous step, the 10 largest values ​​are: 9.8, 9.6, 9.5, 9.3, 9.1, 8.9, 8.7, 8.5, 8.3, and 8.2 Hz / s. The arithmetic mean is (9.8 + 9.6 + 9.5 + 9.3 + 9.1 + 8.9 + 8.7 + 8.5 + 8.3 + 8.2) / 10 = 8.99 Hz / s. Therefore, the peak frequency change rate Rp is taken as 8.99 Hz / s.

[0070] Finally, step S2.4 obtains the disturbance power amplitude stored in step S1, and calculates the system total inertial time constant identification value by combining the grid rated frequency with the frequency change rate peak obtained in S2.3.

[0071] The formula for calculating the total system inertia time constant is derived based on the rotor motion equation. In the initial stage of a disturbance, the relationship between the system frequency change rate and the active power imbalance is: 2H·(dΔf / dt)=ΔP, where H is the total system inertia time constant in seconds; dΔf / dt is the per-unit value of the system frequency change rate, equal to the actual frequency change rate divided by the grid's rated frequency; and ΔP is the per-unit value of the disturbance power amplitude, with the reference capacity being the system's rated capacity.

[0072] Transforming the above equation, we obtain the identified value of the system's total inertia time constant: G = ΔP·f0 / (2·Rp), where G is the identified value of the system's total inertia time constant in seconds; ΔP is the per-unit value of the disturbance power amplitude; f0 is the rated frequency of the power grid in Hz; and Rp is the peak value of the rate of change of frequency in Hz / s. This identified value, in the dimension of time, directly reflects the system's inertia level.

[0073] The per-unit value of the disturbance power amplitude ΔP is calculated as follows: the active value of the disturbance power amplitude stored in step S1 is divided by the system rated capacity baseline value. When the system rated capacity baseline value is taken as the rated power of the energy storage system of 50MW, and the disturbance power amplitude stored in step S1 is 4.8MW, ΔP=4.8 / 50=0.096 per-unit value.

[0074] For example, the grid's rated frequency f0 = 50Hz, the peak rate of frequency change Rp = 8.99Hz / s, and the per-unit disturbance power amplitude ΔP = 0.096. The identified total system inertia time constant G = 0.096 × 50 / (2 × 8.99) = 4.8 / 17.98 ≈ 0.267 seconds. This value indicates that the current system's total inertia level is low, and the corresponding equivalent inertia time constant is approximately 0.267 seconds, which is consistent with the typical characteristics of insufficient system inertia under weak grid conditions.

[0075] It should be noted that the above values ​​are for illustrative purposes only and do not constitute a limitation on the present invention.

[0076] In summary, this step enables online quantitative identification of the total inertia level of the system under the current power grid operating conditions, providing accurate numerical basis for subsequent assessment of inertia shortage and adjustment of energy storage virtual inertia control parameters.

[0077] S3: Based on the current rated frequency of the power grid and the maximum allowable rate of frequency change constraint, calculate the system inertia requirement value required to maintain system frequency security, and determine the system inertia shortage by combining the system total inertia time constant identification value.

[0078] like Figure 2As shown, this step calculates the system inertia demand value by combining the grid rated frequency and the maximum allowable frequency change rate limit, and obtains the initial inertia shortage by subtracting it from the system total inertia time constant identification value. Then, after voltage correction factor amplification and low-pass filtering, the system inertia shortage is obtained.

[0079] Step S3 in the method provided by this invention includes: S3.1: Obtain the maximum allowable rate of change limit specified in the power grid operation specifications, and determine the basic inertia requirement value required to maintain system frequency security in combination with the rated frequency of the power grid; S3.2: Based on the ratio of the peak frequency change rate to the maximum allowable frequency change rate limit, the basic inertia requirement value is proportionally amplified to obtain a system inertia requirement value that is adapted to the current severity of the disturbance; S3.3: Calculate the difference between the system inertia requirement value and the system total inertia time constant identification value to obtain the initial inertia shortage; S3.4: Generate a voltage correction factor based on the per-unit value of the grid connection point voltage amplitude, and fuse the initial inertia shortage with the voltage correction factor to obtain the system inertia shortage, wherein the lower the per-unit value of the grid connection point voltage amplitude, the larger the voltage correction factor; S3.5: Perform low-pass filtering on the system inertia shortage and use the filtered system inertia shortage as the interface quantity with the integral feedback controller in step S4.

[0080] Specifically, step S3.1 first obtains the maximum permissible rate of frequency change limit specified in the power grid operation specifications, and calculates the basic inertia requirement needed to maintain system frequency security in conjunction with the rated frequency of the power grid. The maximum permissible rate of frequency change limit is the constraint value on the rate of frequency change specified in the power grid operation specifications, representing the maximum allowable range of frequency change rate after a disturbance, denoted as R1, with units of Hz / s. Typical values ​​for the maximum permissible rate of frequency change limit are determined according to the operation procedures of different power grids. For example, some power grids specify a maximum permissible rate of frequency change of 0.5 Hz / s, while some power grids with higher frequency quality requirements may specify 0.25 Hz / s.

[0081] The formula for calculating the basic inertia requirement is: G1 = ΔP·f0 / (2·R1), where G1 is the basic inertia requirement in seconds; ΔP is the per-unit value of the disturbance power amplitude; f0 is the rated frequency of the power grid in Hz; and R1 is the maximum allowable rate of frequency change limit in Hz / s. This formula means that under the impact of a disturbance power amplitude ΔP, the minimum inertial time constant required by the system to ensure that the system's rate of frequency change does not exceed the allowable limit R1 is G1.

[0082] For example, if the grid's rated frequency f0 = 50Hz, the maximum allowable rate of change of frequency limit R1 = 0.5Hz / s, the disturbance power amplitude stored in step S1 is 4.8MW, and the system's rated capacity baseline is 50MW, then ΔP = 4.8 / 50 = 0.096. The basic inertia requirement value G1 = 0.096 × 50 / (2 × 0.5) = 4.8 / 1.0 = 4.8 seconds.

[0083] Secondly, step S3.2, based on the ratio of the peak frequency change rate obtained in step S2 to the maximum allowable frequency change rate limit, proportionally amplifies the basic inertia requirement value to obtain a system inertia requirement value that is adapted to the severity of the current disturbance.

[0084] The ratio of the peak frequency change rate R2 to the allowable limit R1 is R2 / R1, reflecting the extent to which the actual frequency change rate caused by the current disturbance exceeds the allowable limit. When the actual peak value is less than or equal to the limit, the ratio is less than or equal to 1, and theoretically, no additional inertia is needed. When the actual peak value exceeds the limit, the ratio is greater than 1, indicating that the existing inertia level cannot limit the frequency change rate within the allowable range, and the basic inertia requirement needs to be increased according to the degree of excess.

[0085] The formula for calculating the system inertia requirement is: G2=G1·(R2 / R1), where G2 is the system inertia requirement in seconds; G1 is the basic inertia requirement in seconds; R2 is the peak frequency change rate in Hz / s; and R1 is the maximum allowable frequency change rate limit in Hz / s.

[0086] For example, the peak frequency change rate obtained in step S2 is R2 = 8.99 Hz / s, the allowable limit is R1 = 0.5 Hz / s, and the basic inertia requirement is G1 = 4.8 seconds. The system inertia requirement is G2 = 4.8 × (8.99 / 0.5) = 4.8 × 17.98 = 86.304 seconds.

[0087] Next, step S3.3 calculates the difference between the system inertia requirement and the identified total system inertia time constant to obtain the initial inertia shortage. The formula for calculating the initial inertia shortage is: G3 = G2 - G0, where G3 is the initial inertia shortage in seconds; G2 is the system inertia requirement in seconds; and G0 is the identified total system inertia time constant obtained in step S2 in seconds. When G3 is greater than 0, it indicates that the current system inertia is insufficient, and energy storage is needed to provide additional virtual inertia compensation; when G3 is less than or equal to 0, it indicates that the current system inertia meets the safety requirements, and no additional compensation is needed.

[0088] For example, the system inertia requirement G2 = 86.304 seconds, and the system total inertia time constant identification value G0 obtained in step S2 = 0.267 seconds. The initial inertia shortage G3 = 86.304 - 0.267 = 86.037 seconds. This value is significantly greater than 0, indicating that the system inertia is severely insufficient under the current weak grid conditions, requiring energy storage to provide a large amount of virtual inertia support.

[0089] Next, step S3.4 generates a voltage correction factor based on the per-unit value of the grid connection point voltage amplitude, and merges the initial inertia shortage with the voltage correction factor to obtain the system inertia shortage.

[0090] The voltage correction factor is generated according to the following rule: the lower the per-unit value of the grid connection point voltage amplitude, the larger the voltage correction factor. This rule means that the deeper the voltage drop at the grid connection point, the more severe the system operating conditions and the higher the risk of frequency instability. Therefore, a greater degree of compensation and amplification is needed based on the initial inertia shortage.

[0091] The voltage correction factor is calculated as follows: K1 = 1 / u, where K1 is the voltage correction factor, which is dimensionless; u is the per-unit value of the grid connection point voltage amplitude, which is dimensionless. When the voltage is normal, u is close to 1, and K1 is close to 1, so it hardly amplifies the initial inertia shortage. When the voltage drops to 0.7, K1 = 1 / 0.7 = 1.4286, and the amplification factor is 1.43 times. When the voltage drops to 0.1, K1 = 10, and the amplification factor is 10 times.

[0092] Meanwhile, to prevent the correction factor from being too large when the voltage is extremely low, causing the inertia shortage to exceed the actual compensable capacity of energy storage, an upper limit is set for the voltage correction factor, with the upper limit being 5. That is, when the per-unit value of the grid connection point voltage amplitude is lower than 0.2, the voltage correction factor remains at 5 and no longer increases as the per-unit value of the voltage amplitude decreases; when the per-unit value of the grid connection point voltage amplitude rises back to 0.2 or above, the voltage correction factor returns to the reciprocal relationship of K1=1 / u.

[0093] The formula for calculating the system inertia shortage is: G4 = G3·K1, where G4 is the system inertia shortage in seconds; G3 is the initial inertia shortage in seconds; and K1 is the voltage correction factor, which is dimensionless.

[0094] For example, in step S1, the voltage drops to 0.70 per unit, and the voltage correction factor K1 = 1 / 0.70 ≈ 1.4286. The initial inertia shortage G3 = 86.037 seconds. The system inertia shortage G4 = 86.037 × 1.4286 ≈ 122.91 seconds.

[0095] Finally, step S3.5 performs low-pass filtering on the system inertia shortage and uses the filtered system inertia shortage as the interface quantity with the integral feedback controller in step S4.

[0096] The purpose of low-pass filtering is to suppress short-time spikes in the system inertia shortage caused by frequency sampling noise, power fluctuation measurement errors, and sudden changes in voltage correction factors, so as to avoid control jitter in the subsequent integral feedback controller due to high-frequency fluctuations in the input.

[0097] A first-order low-pass filter is used, and the cutoff frequency is selected based on the following criteria: the filtered signal should retain the effective change of inertia deficit on a second-scale time scale, while suppressing high-frequency noise on the order of the sampling frequency. The cutoff frequency can be taken as one-tenth of the fundamental frequency of the power grid, i.e., 5Hz in a 50Hz power grid. The time constant corresponding to this cutoff frequency is 1 / (2π×5)≈0.032 seconds. The time required for the filtered output to reach 99% of its steady-state value is approximately 5 times the time constant, i.e., approximately 0.16 seconds. The discretization recursive formula for the first-order low-pass filter is consistent with that in step S1, achieving a 5Hz cutoff frequency by adjusting the time constant. The filtered system inertia deficit is denoted as G5, in seconds.

[0098] For example, the system inertia shortage G4 obtained in step S3.4 is 122.91 seconds. After passing through a first-order low-pass filter with a cutoff frequency of 5Hz, if this value remains stable for several consecutive sampling periods, the filter output G5 converges to around 122.91 seconds in a response time of about 0.16 seconds, which is used as the input for the subsequent step S4.

[0099] It should be noted that the above values ​​are for illustrative purposes only and do not constitute a limitation on the present invention.

[0100] In summary, this step achieves accurate quantification and smooth output of the current grid inertia deficiency, providing a stable and reliable input for the subsequent integral feedback controller.

[0101] S4: Input the system inertia shortage into the integral feedback controller, output the virtual inertia time correction of the energy storage, and limit the virtual inertia time correction of the energy storage converter with the rated power capacity of the energy storage converter as the limiting condition.

[0102] This step involves inputting the system inertia shortage into the integral feedback controller, correcting the integral coefficient by combining it with the available power margin of the energy storage converter, and limiting the output virtual inertia time correction through the limiting unit.

[0103] Step S4 in the method provided by this invention includes: S4.1: Obtain the rated power capacity and default virtual inertia time constant of the energy storage converter, and determine the initial setting value of the integral coefficient based on the rated power capacity; S4.2: Obtain the current output active power of the energy storage converter, combine it with the rated power capacity, obtain the available power margin, and correct the initial setting value of the integral coefficient according to the available power margin to obtain the corrected integral coefficient. S4.3: Determine the output limiting boundary of the integral feedback controller based on the upper and lower limits of the adjustable range of the virtual inertia time constant; S4.4: Using the system inertia shortage as input and the corrected integral coefficient as integral gain, construct an integral operation unit, wherein the output of the integral operation unit is the virtual inertia time correction amount; S4.5: Construct a limiting unit, and connect the output of the integral operation unit to the limiting unit. The limiting unit performs limiting processing on the virtual inertia time correction amount according to the upper limit value and the lower limit value of the output. S4.6: Construct a coefficient adaptive update unit and obtain the ratio of the updated virtual inertia time constant to the current virtual inertia time constant. Adjust the corrected integral coefficients according to the ratio and output the adjusted integral coefficients to the integral operation unit for integral operation in the next control cycle.

[0104] In this step, the integral feedback controller is composed of an integral coefficient correction unit, an integral calculation unit, a limiting unit, and a coefficient adaptive update unit connected together. The integral coefficient correction unit is connected to the rated power capacity, the default virtual inertia time constant, and the current output active power. The integral calculation unit is connected to the system inertia shortage. The limiting unit outputs the limited energy storage virtual inertia time correction. The coefficient adaptive update unit is connected to the updated virtual inertia time constant and feeds back the adjusted integral coefficient to the integral calculation unit.

[0105] Specifically, the integral feedback controller consists of an integral coefficient correction unit, an integral calculation unit, a limiting unit, and a coefficient adaptive update unit connected together. The integral coefficient correction unit receives the rated power capacity, the default virtual inertia time constant, and the current output active power to complete the initial tuning of the integral coefficients and the correction of the available power margin. The integral calculation unit receives the filtered system inertia deficit and performs integral calculations using the corrected integral coefficients as the integral gain. The limiting unit receives the output of the integral calculation unit, performs limiting processing according to the upper and lower limit values, and outputs the limited energy storage virtual inertia time correction. The coefficient adaptive update unit receives the updated virtual inertia time constant, calculates the ratio with the currently effective virtual inertia time constant, adjusts the integral coefficients, and feeds the adjusted integral coefficients back to the integral calculation unit for integral calculations in the next control cycle.

[0106] Specifically, step S4.1 first obtains the rated power capacity and default virtual inertia time constant of the energy storage converter, and determines the initial setting value of the integral coefficient based on the default virtual inertia time constant. The rated power capacity participates in the integral coefficient adjustment in the subsequent available power margin correction.

[0107] The rated power capacity of the energy storage converter is denoted as P0, in MW, and is obtained from the design specifications of the energy storage system's grid connection interface. The default virtual inertia time constant is denoted as T0, in seconds, and is the baseline value of the virtual inertia time constant set at the factory or during initial commissioning of the energy storage system.

[0108] The formula for calculating the initial setting value of the integral coefficient is: K0 = a / T0, where K0 is the initial setting value of the integral coefficient, in seconds. - ¹; 'a' is the tuning coefficient, a dimensionless constant, typically 1; T0 is the default virtual inertia time constant, in seconds. The above formula means that the dimension of the integral coefficient is the reciprocal of the time constant. After multiplying the integral coefficient by the system inertia shortage and the control period, the resulting quantity has the same dimension as the virtual inertia time correction. The rated power capacity does not directly participate in the calculation of the initial integral coefficient tuning value, but rather in the subsequent adjustment of the integral coefficient during the correction of the available power margin.

[0109] For example, the rated power capacity of the energy storage converter is P0 = 50MW, the default virtual inertia time constant is T0 = 4 seconds, and the setting factor is a = 1. The initial setting value of the integral coefficient is K0 = 1 / 4 = 0.25s. - ¹.

[0110] Next, in step S4.2, the current output active power of the energy storage converter is obtained, and the available power margin is calculated in combination with the rated power capacity. The initial setting value of the integral coefficient is then corrected based on the available power margin to obtain the corrected integral coefficient.

[0111] The current output active power is denoted as P1, in MW, and is obtained in real time from the power measurement unit of the energy storage converter. The available power margin is divided into a positive margin P2a and a negative margin P2b. The positive margin P2a = P0 - P1 represents the potential for further increases in active power output from the energy storage converter based on its current output; the negative margin P2b = P0 + P1 represents the potential for further increases in active power absorption from the energy storage converter based on its current output.

[0112] When the system inertia deficit G5 is greater than 0, the integral output increases positively, corresponding to an increase in the discharge power of the energy storage. In this case, a positive margin P2a is selected for correction. When the system inertia deficit G5 is less than 0, the integral output decreases negatively, corresponding to an increase in the charging power of the energy storage. In this case, a negative margin P2b is selected for correction. When the margin in the corresponding direction is less than 10% of the rated power capacity, the integral coefficient is further multiplied by the ratio of the margin in that direction to 10% of the rated power capacity to limit the integral speed in that direction.

[0113] The corrected formula for calculating the integral coefficient is as follows: when G5 is greater than 0, K1 = K0·(P2a / P0); when G5 is less than 0, K1 = K0·(P2b / P0). Where K1 is the corrected integral coefficient, in seconds. - ¹; K0 is the initial setting value of the integral coefficient, in seconds. - ¹; P2a is the positive margin in MW; P2b is the negative margin in MW; P0 is the rated power capacity in MW. The above correction means that when the current output active power of the energy storage converter is high, the available power margin is small, and the corrected integral coefficient is reduced accordingly to avoid the integral output from growing too fast and causing the power to exceed the limit; when the available power margin is sufficient, the integral coefficient is close to the initial setting value.

[0114] For example, the current output active power P1 = 30MW, the rated power capacity P0 = 50MW, the positive margin P2a = 50 - 30 = 20MW, and the negative margin P2b = 50 + 30 = 80MW. The corrected integral coefficient K1 = 0.25 × (20 / 50) = 0.25 × 0.4 = 0.1s - ¹.

[0115] Next, step S4.3 determines the output limiting boundary of the integral feedback controller based on the upper and lower limits of the adjustable range of the virtual inertia time constant. The upper limit of the adjustable range of the virtual inertia time constant is denoted as Tmax, and the lower limit is denoted as Tmin, both in seconds. This range is determined by the control capability constraints of the energy storage converter and the grid operation requirements. The output of the integral feedback controller is the virtual inertia time correction, which physically represents the increment based on the default virtual inertia time constant. Therefore, the output limiting boundary is determined by the difference between the upper and lower limits of the adjustable range and the default virtual inertia time constant.

[0116] The upper limit of the output amplitude is denoted as Lmax, and the lower limit of the output amplitude is denoted as Lmin, both in seconds. The calculation formula is: Lmax=Tmax-T0, Lmin=Tmin-T0, where Tmax is the upper limit of the adjustable range of the virtual inertia time constant, Tmin is the lower limit, and T0 is the default virtual inertia time constant.

[0117] For example, the adjustable range of the virtual inertia time constant is Tmax = 10 seconds and Tmin = 1 second, with a default virtual inertia time constant T0 = 4 seconds. The upper limit of the output amplitude is Lmax = 10 - 4 = 6 seconds, and the lower limit of the output amplitude is Lmin = 1 - 4 = -3 seconds.

[0118] Next, step S4.4 uses the system inertia shortage as input and the corrected integral coefficient as the integral gain to construct an integral operation unit. The output of the integral operation unit is the virtual inertia time correction. The discretized recursive formula of the integral operation unit is: ΔT(k) = ΔT(k-1) + K1·G5(k)·Ts, where ΔT(k) is the virtual inertia time correction for the current control cycle in seconds; ΔT(k-1) is the virtual inertia time correction for the previous control cycle in seconds; and K1 is the corrected integral coefficient in seconds. - ¹; G5(k) is the system inertia deficit after filtering in the current control cycle, in seconds; Ts is the control cycle, in seconds. The dimension of K1·G5(k)·Ts is (s) - ¹) × seconds × seconds = seconds, consistent with ΔT(k).

[0119] The function of the integral operation unit is as follows: when the system inertia shortage is positive, the integral output continues to increase, causing the virtual inertia time correction to increase positively; when the system inertia shortage is negative, the integral output continues to decrease, causing the virtual inertia time correction to decrease negatively; when the system inertia shortage is zero, the integral output remains unchanged.

[0120] For example, the virtual inertia time correction ΔT(k-1) = 1.5 seconds in the previous control cycle, and the corrected integral coefficient K1 = 0.1 seconds. - ¹, The filtered system inertia shortage G5 = 122.91 seconds, and the control period Ts = 0.001 seconds. The virtual inertia time correction ΔT(k) for the current control period = 1.5 + 0.1 × 122.91 × 0.001 = 1.5 + 0.012291 = 1.512291 seconds.

[0121] Further, in step S4.5, a limiting unit is constructed, and the output of the integration unit is connected to the limiting unit. The limiting unit limits the virtual inertia time correction according to the upper and lower limits of the output amplitude.

[0122] The limiting processing rules are as follows: when the virtual inertia time correction output by the integration unit is greater than the upper limit value Lmax, the limiting unit outputs Lmax; when the virtual inertia time correction is less than the lower limit value Lmin, the limiting unit outputs Lmin; when the virtual inertia time correction is between the upper and lower limits, the limiting unit outputs the original value.

[0123] For example, the virtual inertia time correction ΔT(k) output by the integration unit is 1.512291 seconds, the upper limit amplitude Lmax is 6 seconds, and the lower limit amplitude Lmin is -3 seconds. 1.512291 seconds is between -3 seconds and 6 seconds, and the limiting unit outputs the original value of 1.512291 seconds.

[0124] Finally, in step S4.6, an adaptive coefficient update unit is constructed to obtain the ratio of the updated virtual inertia time constant to the current virtual inertia time constant. The corrected integral coefficients are adjusted according to the ratio, and the adjusted integral coefficients are output to the integral calculation unit for integral calculation in the next control cycle.

[0125] The updated virtual inertia time constant is denoted as T1, in seconds, and is calculated in step S5 based on the virtual inertia time correction after limiting. The current virtual inertia time constant is the virtual inertia time constant that is actually effective in this control cycle.

[0126] The adjustment formula for the coefficient adaptive update unit is: K2 = K1·(T1 / T2), where K2 is the adjusted integral coefficient in seconds. - ¹; K1 is the corrected integral coefficient, in units of s. - ¹; T1 is the updated virtual inertia time constant in seconds; T2 is the actual effective virtual inertia time constant in the current control cycle in seconds. The above adjustments mean that when the virtual inertia time constant increases, the integral coefficient increases accordingly to match the response speed of subsequent integral operations with the new virtual inertia time constant level; when the virtual inertia time constant decreases, the integral coefficient decreases accordingly to avoid control oscillations caused by excessively rapid changes in the integral output.

[0127] In the first control cycle, both the updated virtual inertia time constant T1 and the actual effective virtual inertia time constant T2 in the current control cycle are taken as the default virtual inertia time constant T0. The ratio T1 / T2 output by the coefficient adaptive update unit is 1, and the integral coefficient in the first control cycle is not adaptively adjusted. From the second control cycle onwards, T1 is taken as the updated virtual inertia time constant output in step S5 of the previous control cycle, and T2 is taken as the actual effective virtual inertia time constant of the previous control cycle.

[0128] For example, in the current control cycle, the actual effective virtual inertia time constant T2 = 4 seconds. The virtual inertia time correction after limiting is 1.512291 seconds. After updating in step S5, the virtual inertia time constant T1 = 4 + 1.512291 = 5.512291 seconds. The adjusted integral coefficient K2 = 0.1 × (5.512291 / 4) = 0.1 × 1.378 = 0.1378s -¹. The adjusted integral coefficient is output to the integral calculation unit for integral calculation in the next control cycle.

[0129] It should be noted that the above values ​​are for illustrative purposes only and do not constitute a limitation on the present invention.

[0130] In summary, this step achieves adaptive calculation and safety limiting of the virtual inertia time correction for energy storage, ensuring that the correction is within the power range that the energy storage converter can withstand, and providing a reliable adjustment amount for subsequent updates to the virtual inertia time constant and control parameters.

[0131] S5: Update the virtual inertia time constant of the energy storage according to the virtual inertia time correction after the limit, and update the droop coefficient synchronously based on the virtual inertia time constant before and after the update to obtain the updated control parameter combination.

[0132] This step involves summing the virtual inertia time correction after limiting with the current virtual inertia time constant and then limiting it to obtain the updated virtual inertia time constant. Simultaneously, the voltage support bias factor is calculated based on the ratio of the virtual inertia time constant before and after the update and the per-unit value of the grid connection point voltage amplitude. The droop coefficient is then adjusted synchronously to obtain the updated control parameter combination.

[0133] Step S5 in the method provided by this invention includes: S5.1: Sum the current virtual inertia time constant with the limited energy storage virtual inertia time correction amount to obtain the updated virtual inertia time constant; S5.2: Compare the updated virtual inertia time constant with the upper and lower limits of the adjustable range of the virtual inertia time constant. If it is determined that the virtual inertia time constant exceeds the upper limit or falls below the lower limit, the upper limit or the lower limit is used as the updated virtual inertia time constant, respectively; otherwise, it remains unchanged. S5.3: Calculate the ratio of the updated virtual inertia time constant to the current virtual inertia time constant, and use it as the droop coefficient synchronous adjustment scaling factor; S5.4: Calculate the voltage support bias factor based on the per-unit value of the voltage amplitude at the grid connection point; S5.5: Determine the updated droop coefficient based on the current droop coefficient, the droop coefficient synchronous adjustment ratio factor, and the voltage support bias factor; S5.6: Combine the updated virtual inertia time constant after the limiting process in S5.2 with the updated droop coefficient to form an optimized control parameter group, and replace the current virtual inertia time constant and the current droop coefficient in the self-synchronizing voltage source control loop of the energy storage converter.

[0134] Specifically, step S5.1 first sums the current virtual inertia time constant with the limited energy storage virtual inertia time correction to obtain the updated virtual inertia time constant. The current virtual inertia time constant is denoted as T2, in seconds, and is the actual effective virtual inertia time constant in this control cycle. The limited energy storage virtual inertia time correction is denoted as ΔT, in seconds, and is output by the limiting unit in step S4. The updated virtual inertia time constant is calculated as: T1 = T2 + ΔT, where T1 is the updated virtual inertia time constant in seconds.

[0135] For example, in the current control cycle, the actual effective virtual inertia time constant T2 = 4 seconds, and the virtual inertia time correction amount ΔT output by the limiting unit in step S4 = 1.512291 seconds. The updated virtual inertia time constant T1 = 4 + 1.512291 = 5.512291 seconds.

[0136] Next, step S5.2 compares the updated virtual inertia time constant with the upper and lower limits of the adjustable range of the virtual inertia time constant. When the updated virtual inertia time constant exceeds the upper limit, the upper limit is used as the updated virtual inertia time constant; when it is lower than the lower limit, the lower limit is used as the updated virtual inertia time constant; when it is between the upper and lower limits, the original value remains unchanged.

[0137] The upper limit of the adjustable range for the virtual inertia time constant is Tmax, and the lower limit is Tmin, both in seconds. The limiting rule is: when T1 is greater than Tmax, T1 = Tmax; when T1 is less than Tmin, T1 = Tmin; otherwise, T1 remains unchanged.

[0138] For example, the updated virtual inertia time constant T1 = 5.512291 seconds, the upper limit Tmax = 10 seconds, and the lower limit Tmin = 1 second. 5.512291 seconds is between 1 second and 10 seconds, and T1 remains at 5.512291 seconds after limiting.

[0139] Next, step S5.3 calculates the ratio of the updated virtual inertia time constant to the current virtual inertia time constant, which is used as the droop coefficient synchronization adjustment scaling factor. The calculation formula for the droop coefficient synchronization adjustment scaling factor is: K3=T1 / T2, where K3 is the droop coefficient synchronization adjustment scaling factor, which is dimensionless; T1 is the updated virtual inertia time constant after the limiting processing in S5.2, in seconds; and T2 is the current virtual inertia time constant, in seconds.

[0140] For example, after the clipping is limited, the updated virtual inertia time constant T1 = 5.512291 seconds, and the current virtual inertia time constant T2 = 4 seconds. The droop coefficient synchronous adjustment scaling factor K3 = 5.512291 / 4 = 1.378.

[0141] Next, step S5.4 calculates the voltage support bias factor based on the per-unit value of the grid connection point voltage amplitude. The voltage support bias factor is inversely proportional to the per-unit value of the grid connection point voltage amplitude; that is, the deeper the voltage drop, the larger the voltage support bias factor.

[0142] In this invention, the voltage correction factor is used to correct the inertia shortage, enhancing the inertia compensation under weak grid conditions; the voltage support bias factor is used to correct the droop coefficient, adjusting the strength of reactive voltage support under weak grid conditions. Both are constructed based on the per-unit value of the grid connection point voltage amplitude, but they act on the inertia compensation stage and the droop adjustment stage respectively, with independent functions and different applicable scenarios. The voltage correction factor acts on the input of the integral feedback controller, increasing the virtual inertia time correction by amplifying the inertia shortage; the voltage support bias factor acts on the droop coefficient, enhancing or suppressing reactive voltage support capability by amplifying the droop coefficient.

[0143] The voltage correction factor is in reciprocal form because the inertia shortage, as an input to the integral feedback controller, is highly sensitive to voltage drops and requires strong nonlinear amplification to rapidly increase the virtual inertia compensation during deep voltage drops. The voltage support bias factor uses a dead-zone plus linear deviation form because the adjustable range of the droop coefficient is relatively limited; a linear approach is easier to coordinate with the upper limit of the droop coefficient, preventing excessively large step changes in the droop coefficient during voltage drops.

[0144] In this step, step S5.4 includes: S5.4.1: A dead zone interval is constructed with the rated voltage per unit value as a reference point. When the voltage amplitude per unit value at the grid connection point is within the dead zone interval, the voltage support bias factor is set to 1. The dead zone interval is centered on the rated voltage per unit value. The upper limit is the sum of the rated voltage per unit value and the allowable positive voltage deviation value in the grid operation specification. The lower limit is the difference between the rated voltage per unit value and the allowable negative voltage deviation value in the grid operation specification. The closed interval between the upper limit and the lower limit constitutes the dead zone interval. S5.4.2: When the per-unit value of the grid connection point voltage amplitude is lower than the lower limit of the dead zone, the difference between the per-unit value of the grid connection point voltage amplitude and the lower limit of the dead zone is calculated as the voltage deviation, and the voltage support bias factor is the sum of 1 and the voltage deviation. S5.4.3: When the per-unit value of the grid connection point voltage amplitude is higher than the upper limit of the dead zone interval, the difference between the upper limit of the dead zone interval and the per-unit value of the grid connection point voltage amplitude is calculated as the voltage over-limit amount. The voltage support bias factor is the difference between 1 and the voltage over-limit amount. If the calculation result is negative, the voltage support bias factor is 0.

[0145] Specifically, step S5.4.1 constructs a dead zone interval using the rated voltage per-unit value as a reference point. The rated voltage per-unit value is denoted as u0, and under ideal operating conditions, u0 = 1.0. The dead zone interval is centered on the rated voltage per-unit value, with the upper limit denoted as u2 and the lower limit denoted as u1. The upper limit is the sum of the rated voltage per-unit value and the allowable positive voltage deviation value in the power grid operation specification, and the lower limit is the difference between the rated voltage per-unit value and the allowable negative voltage deviation value in the power grid operation specification.

[0146] In power grid operation specifications, the permissible positive voltage deviation is denoted as e2, and the permissible negative voltage deviation is denoted as e1, both of which are dimensionless per-unit values. These permissible positive and negative voltage deviations are obtained by referencing the voltage operation management regulations issued by the power grid dispatching agency. For example, if the permissible positive voltage deviation is +5%, then e2 = 0.05; and the permissible negative voltage deviation is -3%, then e1 = -0.03. The corresponding permissible voltage range is 0.97 to 1.05 per unit.

[0147] The formulas for calculating the upper and lower limits of the dead zone are: u2 = u0 + e2, u1 = u0 + e1, where u2 is the upper limit of the dead zone, u1 is the lower limit of the dead zone, u0 is the per-unit value of the rated voltage, e2 is the allowable positive voltage deviation, and e1 is the allowable negative voltage deviation. The closed interval between the upper and lower limits constitutes the dead zone.

[0148] When the per-unit voltage amplitude u at the grid connection point is within the dead zone (i.e., u1≤u≤u2), the voltage support bias factor K4 is set to 1. This design means that when the voltage deviation is within the allowable range of the power grid operation specifications, the voltage support bias factor does not provide additional correction, and the droop coefficient is only affected by the droop coefficient synchronous adjustment proportional factor.

[0149] For example, the rated voltage per unit value u0 = 1.0, the allowable positive voltage deviation e2 = 0.05, and the allowable negative voltage deviation e1 = -0.03. The upper limit of the dead zone u2 = 1.0 + 0.05 = 1.05, and the lower limit u1 = 1.0 - 0.03 = 0.97. The dead zone is [0.97, 1.05]. If the per-unit voltage amplitude at the grid connection point u = 0.98, which is within the dead zone, K4 is set to 1.

[0150] Further, in step S5.4.2, when the per-unit value of the grid connection point voltage amplitude is lower than the lower limit of the dead zone interval, the difference between the per-unit value of the grid connection point voltage amplitude and the lower limit of the dead zone interval is calculated as the voltage deviation, and the voltage support bias factor is the sum of 1 and the voltage deviation.

[0151] The formula for calculating voltage deviation is: Δu1 = u1 - u, where Δu1 is the voltage deviation, dimensionless and always positive; u is the per-unit value of the voltage amplitude at the grid connection point; and u1 is the lower limit of the dead zone. The formula for calculating the voltage support bias factor is: K4 = 1 + Δu1, where K4 is the voltage support bias factor, dimensionless. When the voltage is below the lower limit of the dead zone, K4 is greater than 1. The deeper the voltage drop, the larger Δu1 and K4 become, the more significant the droop factor amplification, and the stronger the reactive power voltage support capability.

[0152] To prevent K4 from becoming too large during voltage dips, an upper limit is set for the voltage support bias factor, with the upper limit set to 2. When the calculated K4 is greater than 2, K4 is set to 2. Under the condition of a dead zone lower limit of u1 = 0.97, when the per-unit value of the grid connection point voltage amplitude drops to 0, K4 reaches its maximum value of 1.97, which does not exceed the upper limit of 2. Therefore, this upper limit will never be triggered in actual operation and is only used as a protective boundary under extreme conditions.

[0153] For example, the per-unit value of the grid connection point voltage amplitude is u=0.70, and the lower limit of the dead zone is u1=0.97. The voltage deviation Δu1=0.97-0.70=0.27. The voltage support bias factor K4=1+0.27=1.27. Since it does not exceed the upper limit of 2, K4 is set to 1.27.

[0154] Further, in step S5.4.3, when the per-unit value of the grid connection point voltage amplitude is higher than the upper limit of the dead zone interval, the difference between the upper limit of the dead zone interval and the per-unit value of the grid connection point voltage amplitude is calculated as the voltage over-limit amount, and the voltage support bias factor is the difference between 1 and the voltage over-limit amount. If the calculation result is negative, the voltage support bias factor is set to 0.

[0155] The formula for calculating the voltage over-limit is: Δu2 = u2 - u, where Δu2 is the voltage over-limit, dimensionless, and always negative; u is the per-unit value of the voltage amplitude at the grid connection point; and u2 is the upper limit of the dead zone. The formula for calculating the voltage support bias factor is: K4 = 1 + Δu2, which is equivalent to K4 = 1 - (u - u2), meaning that when the voltage is higher than the upper limit of the dead zone, K4 is less than 1. The higher the voltage, the smaller K4, and the corresponding decrease in the droop coefficient, thus weakening the reactive power support strength under high voltage. When the calculated result of K4 is negative, K4 is set to 0.

[0156] For example, the per-unit voltage amplitude at the grid connection point is u=1.12, and the upper limit of the dead zone is u2=1.05. The voltage overshoot Δu2=1.05-1.12=-0.07. The voltage support bias factor K4=1-0.07=0.93, which is positive, so K4 is set to 0.93. When the per-unit voltage amplitude at the grid connection point is u=1.25, the voltage overshoot Δu2=1.05-1.25=-0.20, and K4=1-0.20=0.80, which is still positive. When u=2.10, Δu2=1.05-2.10=-1.05, and K4=1-1.05=-0.05, which is negative, so K4 is set to 0.

[0157] Further, step S5.5 determines the updated droop coefficient based on the current droop coefficient, the droop coefficient synchronization adjustment factor, and the voltage support bias factor. The updated droop coefficient is calculated using the formula: D1 = D0·K3·K4, where D1 is the updated droop coefficient (dimensionless); D0 is the current droop coefficient (dimensionless); K3 is the droop coefficient synchronization adjustment factor (dimensionless); and K4 is the voltage support bias factor (dimensionless).

[0158] To prevent the droop coefficient from becoming too large, an upper limit is set for the updated droop coefficient, which is three times the current droop coefficient. When the calculation result exceeds the upper limit, the upper limit is used as the updated droop coefficient.

[0159] For example, the current droop coefficient D0 = 0.05, the droop coefficient synchronization adjustment ratio factor K3 = 1.378, and the voltage support bias factor K4 = 1.27. The updated droop coefficient D1 = 0.05 × 1.378 × 1.27 ≈ 0.0875. The upper limit is 0.05 × 3 = 0.15. 0.0875 does not exceed the upper limit, so the updated droop coefficient is taken as 0.0875.

[0160] Finally, step S5.6 combines the updated virtual inertia time constant after the limiting process in S5.2 with the updated droop coefficient into an optimized control parameter group, replacing the current virtual inertia time constant and the current droop coefficient in the self-synchronizing voltage source control loop of the energy storage converter.

[0161] The optimized control parameter set includes two parameters: the updated virtual inertia time constant T1 and the updated droop coefficient D1. These two parameters are written into the corresponding registers or control variables of the energy storage converter's self-synchronizing voltage source control loop, replacing the virtual inertia time constant T2 and the current droop coefficient D0 that are effective in the current control cycle. From the next control cycle onwards, the virtual inertia response and reactive voltage support characteristics of the energy storage converter are executed according to the updated parameters.

[0162] For example, after being limited by S5.2, the updated virtual inertia time constant T1 = 5.512291 seconds, and the updated droop coefficient D1 = 0.0875. These two parameters are written into the control loop, replacing the current virtual inertia time constant T2 = 4 seconds and the current droop coefficient D0 = 0.05. From the next control cycle onwards, it will be executed with T1 = 5.512291 seconds and D1 = 0.0875.

[0163] It should be noted that the above values ​​are for illustrative purposes only and do not constitute a limitation on the present invention.

[0164] In summary, this step achieves coordinated and adaptive updates of virtual inertia parameters and droop coefficients, improving the system's inertia support capability while also taking into account reactive power and voltage support requirements, thus providing parameter assurance for the stable operation of the energy storage converter under weak grid conditions.

[0165] In summary, this invention can adjust the energy storage control parameters in real time under the conditions of power grid strength changes and voltage drops, which not only makes up for the insufficient system inertia, but also enhances the reactive voltage support capability under weak power grid conditions, effectively improving the frequency stability and operational adaptability of the energy storage converter under weak power grid conditions.

[0166] Example 2, as Figure 3 As shown, the present invention provides a parameter optimization and tuning system for a self-synchronizing voltage source energy storage controller, the system comprising: The event detection and data capture module 11 is used to continuously monitor the active power fluctuation amplitude of the energy storage grid connection point. When the active power fluctuation amplitude exceeds the trigger threshold, the current time is recorded as the trigger time, and the frequency change curve of the grid connection point within a preset time period is captured from the trigger time.

[0167] Specifically, the active power fluctuation amplitude of the energy storage grid connection point is continuously monitored. When the active power fluctuation amplitude exceeds the trigger threshold, the current time is recorded as the trigger time, and the frequency change curve of the grid connection point within a preset time period is extracted from the trigger time, including: The fundamental components of the instantaneous values ​​of the three-phase voltage and three-phase current at the energy storage grid connection point are extracted to obtain the fundamental value of the active power. The root mean square value of the active power fluctuation amplitude is calculated using a sliding time window and used as the real-time active power fluctuation amplitude. Obtain the per-unit value of the voltage amplitude at the grid connection point, and multiply the reference trigger threshold by the per-unit value of the voltage amplitude to obtain the trigger threshold that is adaptively adjusted according to the voltage drop depth; When the real-time active power fluctuation exceeds the adaptive adjustment trigger threshold, the current time is recorded as the trigger time, and the real-time active power fluctuation is stored as the disturbance power amplitude. Starting from the triggering time, a grid connection point frequency change curve with a length of two power frequency cycles is extracted. The grid connection point frequency change curve is composed of the instantaneous frequency values ​​output by the phase-locked loop arranged continuously according to the sampling points.

[0168] The step of obtaining the per-unit value of the grid connection point voltage amplitude includes: Input the instantaneous three-phase voltage values ​​at the energy storage grid connection point into the Parker transformation matrix, and output the d-axis voltage components and q-axis voltage components in the rotating coordinate system; The d-axis voltage component and the q-axis voltage component are filtered by a first-order low-pass filter with a cutoff frequency of twice the fundamental frequency, respectively, to obtain the filtered d-axis voltage component and the filtered q-axis voltage component. Calculate the sum of squares of the filtered d-axis voltage component and the filtered q-axis voltage component, and take the square root of the sum of squares to obtain the fundamental amplitude of the grid connection point voltage. A moving average window is constructed with half a power frequency cycle as the window length. The fundamental amplitude of the grid connection point voltage is processed by moving average, and the result after moving average processing is used as the smoothed voltage amplitude. Divide the smoothed voltage amplitude by the AC voltage reference value to obtain the per-unit value of the grid connection point voltage amplitude, wherein the AC voltage reference value is the phase voltage amplitude reference value corresponding to the rated line voltage of the grid connection point.

[0169] The total inertia online identification module 12 is used to calculate the system total inertia time constant identification value under the current power grid operating state based on the peak value of the frequency change rate of the frequency change curve at the grid connection point within the first power frequency cycle after the triggering time, combined with the active power fluctuation amplitude.

[0170] Specifically, based on the peak value of the frequency change rate of the frequency change curve at the grid connection point within the first power frequency cycle after the triggering time, and combined with the active power fluctuation amplitude, the system total inertia time constant identification value under the current grid operating state is calculated, including: Extract the frequency sampling point sequence within the first power frequency cycle after the trigger moment from the frequency change curve of the grid connection point, and arrange them in chronological order of sampling time; Calculate the absolute value of the difference between instantaneous frequency values ​​between adjacent sampling points in the frequency sampling point sequence to obtain a sequence of absolute values ​​of frequency change rate; The values ​​in the absolute value sequence of frequency change rate are sorted in descending order of magnitude, and the arithmetic mean of the first N values ​​after sorting is taken as the peak value of frequency change rate, where N is determined according to the total number of sampling points in the first power frequency cycle. The disturbance power amplitude is obtained, and combined with the grid rated frequency and the peak value of the frequency change rate, the system total inertia time constant identification value in terms of time is obtained.

[0171] The inertia shortage assessment module 13 is used to calculate the system inertia demand value required to maintain system frequency security based on the current grid rated frequency and the maximum allowable frequency change rate constraint, and to determine the system inertia shortage by combining the system total inertia time constant identification value.

[0172] Specifically, based on the current grid rated frequency and the maximum allowable rate of frequency change constraint, the system inertia requirement for maintaining system frequency security is calculated, and combined with the identified total system inertia time constant, the system inertia shortage is determined, including: Obtain the maximum allowable rate of change of frequency limit specified in the power grid operation specifications, and determine the basic inertia requirement value required to maintain system frequency security in combination with the rated frequency of the power grid; Based on the ratio of the peak frequency change rate to the maximum allowable frequency change rate limit, the basic inertia requirement value is proportionally amplified to obtain a system inertia requirement value that is adapted to the severity of the current disturbance. The initial inertia shortage is obtained by subtracting the system inertia requirement value from the system total inertia time constant identification value. A voltage correction factor is generated based on the per-unit value of the grid connection point voltage amplitude, and the initial inertia shortage is fused with the voltage correction factor to obtain the system inertia shortage. The lower the per-unit value of the grid connection point voltage amplitude, the larger the voltage correction factor. The system inertia shortage is subjected to low-pass filtering, and the filtered system inertia shortage is used as the interface quantity with the integral feedback controller.

[0173] The correction calculation and limiting module 14 is used to input the system inertia shortage into the integral feedback controller, output the virtual inertia time correction of the energy storage, and limit the virtual inertia time correction with the rated power capacity of the energy storage converter as the limiting condition.

[0174] Specifically, the system inertia shortage is input into the integral feedback controller, which outputs a virtual inertia time correction value for energy storage. This virtual inertia time correction value is then limited using the rated power capacity of the energy storage converter as a limiting condition, including: Obtain the rated power capacity and default virtual inertia time constant of the energy storage converter, and determine the initial setting value of the integral coefficient based on the rated power capacity; Obtain the current output active power of the energy storage converter, combine it with the rated power capacity to obtain the available power margin, and correct the initial setting value of the integral coefficient according to the available power margin to obtain the corrected integral coefficient. The output limiting boundary of the integral feedback controller is determined based on the upper and lower limits of the adjustable range of the virtual inertia time constant. Using the system inertia shortage as input and the corrected integral coefficient as integral gain, an integral operation unit is constructed, wherein the output of the integral operation unit is the virtual inertia time correction amount; A limiting unit is constructed, and the output of the integral operation unit is connected to the limiting unit. The limiting unit performs limiting processing on the virtual inertia time correction amount according to the upper limit value and the lower limit value of the output. An adaptive coefficient update unit is constructed to obtain the ratio of the updated virtual inertia time constant to the current virtual inertia time constant. The corrected integral coefficient is adjusted according to the ratio, and the adjusted integral coefficient is output to the integral calculation unit for integral calculation in the next control cycle.

[0175] The integral feedback controller is composed of an integral coefficient correction unit, an integral calculation unit, a limiting unit, and a coefficient adaptive update unit. The integral coefficient correction unit is connected to the rated power capacity, the default virtual inertia time constant, and the current output active power. The integral calculation unit is connected to the system inertia shortage. The limiting unit outputs the limited energy storage virtual inertia time correction. The coefficient adaptive update unit is connected to the updated virtual inertia time constant and feeds back the adjusted integral coefficient to the integral calculation unit.

[0176] The control parameter update module 15 is used to update the virtual inertia time constant of the energy storage according to the virtual inertia time correction after the limit, and to update the droop coefficient synchronously based on the virtual inertia time constant before and after the update, so as to obtain the updated control parameter combination.

[0177] Specifically, the virtual inertia time constant of the energy storage is updated based on the virtual inertia time correction after the limit is set, and the droop coefficient is updated synchronously based on the virtual inertia time constant before and after the update to obtain the updated control parameter combination, including: The current virtual inertia time constant is summed with the limited energy storage virtual inertia time correction amount to obtain the updated virtual inertia time constant; The updated virtual inertia time constant is compared with the upper and lower limits of the adjustable range of the virtual inertia time constant. If it is determined that the virtual inertia time constant exceeds the upper limit or falls below the lower limit, the upper limit or the lower limit is used as the updated virtual inertia time constant, respectively; otherwise, it remains unchanged. Calculate the ratio of the updated virtual inertia time constant to the current virtual inertia time constant, and use it as the scaling factor for synchronous adjustment of the droop coefficient; Calculate the voltage support bias factor based on the per-unit value of the voltage amplitude at the grid connection point; Based on the current droop coefficient, the droop coefficient synchronous adjustment ratio factor, and the voltage support bias factor, the updated droop coefficient is determined. The updated virtual inertia time constant after the amplitude limiting process is combined with the updated droop coefficient to form an optimized control parameter group, which replaces the current virtual inertia time constant and the current droop coefficient in the self-synchronizing voltage source control loop of the energy storage converter.

[0178] The calculation of the voltage support bias factor based on the per-unit value of the grid connection point voltage amplitude includes: A dead zone interval is constructed with the rated voltage per unit value as a reference point. When the voltage amplitude per unit value at the grid connection point is within the dead zone interval, the voltage support bias factor is set to 1. The dead zone interval is centered on the rated voltage per unit value. The upper limit is the sum of the rated voltage per unit value and the allowable positive voltage deviation value in the grid operation specification. The lower limit is the difference between the rated voltage per unit value and the allowable negative voltage deviation value in the grid operation specification. The closed interval between the upper limit and the lower limit constitutes the dead zone interval. When the per-unit value of the grid connection point voltage amplitude is lower than the lower limit of the dead zone, the difference between the per-unit value of the grid connection point voltage amplitude and the lower limit of the dead zone is calculated as the voltage deviation. The voltage support bias factor is the sum of 1 and the voltage deviation. When the per-unit value of the grid connection point voltage amplitude is higher than the upper limit of the dead zone, the difference between the upper limit of the dead zone and the per-unit value of the grid connection point voltage amplitude is calculated as the voltage over-limit amount. The voltage support bias factor is the difference between 1 and the voltage over-limit amount. If the calculation result is negative, the voltage support bias factor is 0.

[0179] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0180] This specification and accompanying drawings are merely illustrative examples of the invention and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of the invention. Clearly, those skilled in the art can make various alterations and modifications to the invention without departing from its scope. Therefore, if such modifications and modifications fall within the scope of the invention and its equivalents, the invention is intended to include these modifications and modifications.

Claims

1. A method for optimizing and tuning parameters of a self-synchronizing voltage source energy storage controller, characterized in that, The method includes: S1: Continuously monitor the active power fluctuation amplitude of the energy storage grid connection point. When the active power fluctuation amplitude exceeds the trigger threshold, record the current time as the trigger time, and extract the grid connection point frequency change curve within a preset time period starting from the trigger time. S2: Based on the peak value of the frequency change rate of the frequency change curve at the grid connection point within the first power frequency cycle after the triggering time, and combined with the active power fluctuation amplitude, calculate the identification value of the total system inertia time constant under the current power grid operating state; S3: Based on the current rated frequency of the power grid and the maximum allowable rate of frequency change constraint, calculate the system inertia requirement value required to maintain system frequency security, and determine the system inertia shortage by combining the system total inertia time constant identification value; S4: Input the system inertia shortage into the integral feedback controller, output the virtual inertia time correction of the energy storage, and limit the virtual inertia time correction of the energy storage converter with the rated power capacity of the energy storage converter as the limiting condition; S5: Update the virtual inertia time constant of the energy storage according to the virtual inertia time correction after the limit, and update the droop coefficient synchronously based on the virtual inertia time constant before and after the update to obtain the updated control parameter combination.

2. The method for optimizing and tuning parameters of a self-synchronizing voltage source energy storage controller according to claim 1, characterized in that, Step S1 includes: S1.1: Extract the fundamental components of the instantaneous values ​​of the three-phase voltage and three-phase current at the energy storage grid connection point, obtain the fundamental value of active power, and use a sliding time window to calculate the root mean square value of the active power fluctuation amplitude as the real-time active power fluctuation amplitude. S1.2: Obtain the per-unit value of the voltage amplitude at the grid connection point, multiply the reference trigger threshold by the per-unit value of the voltage amplitude, and obtain the trigger threshold that is adaptively adjusted according to the voltage drop depth; S1.3: When the real-time active power fluctuation amplitude exceeds the adaptive adjustment trigger threshold, the current time is recorded as the trigger time, and the real-time active power fluctuation amplitude is stored as the disturbance power amplitude. S1.4: Starting from the triggering time, extract the grid connection point frequency change curve with a length of two power frequency cycles, wherein the grid connection point frequency change curve is composed of the instantaneous frequency values ​​output by the phase-locked loop arranged continuously according to the sampling points.

3. The method for optimizing and tuning parameters of a self-synchronizing voltage source energy storage controller according to claim 2, characterized in that, The steps for obtaining the per-unit value of the grid connection point voltage amplitude include: Input the instantaneous three-phase voltage values ​​at the energy storage grid connection point into the Parker transformation matrix, and output the d-axis voltage components and q-axis voltage components in the rotating coordinate system; The d-axis voltage component and the q-axis voltage component are filtered by a first-order low-pass filter with a cutoff frequency of twice the fundamental frequency, respectively, to obtain the filtered d-axis voltage component and the filtered q-axis voltage component. Calculate the sum of squares of the filtered d-axis voltage component and the filtered q-axis voltage component, and take the square root of the sum of squares to obtain the fundamental amplitude of the grid connection point voltage. A moving average window is constructed with half a power frequency cycle as the window length. The fundamental amplitude of the grid connection point voltage is processed by moving average, and the result after moving average processing is used as the smoothed voltage amplitude. Divide the smoothed voltage amplitude by the AC voltage reference value to obtain the per-unit value of the grid connection point voltage amplitude, wherein the AC voltage reference value is the phase voltage amplitude reference value corresponding to the rated line voltage of the grid connection point.

4. The method for optimizing and tuning parameters of a self-synchronizing voltage source energy storage controller according to claim 1, characterized in that, Step S2 includes: S2.1: Extract the frequency sampling point sequence within the first power frequency cycle after the triggering time from the frequency change curve of the grid connection point, and arrange them in order of sampling time; S2.2: Calculate the absolute value of the difference between instantaneous frequency values ​​between adjacent sampling points in the frequency sampling point sequence to obtain a sequence of absolute values ​​of frequency change rate; S2.3: Sort each value in the absolute value sequence of frequency change rate in descending order, and take the arithmetic mean of the first N values ​​after sorting as the peak value of frequency change rate, where N is determined according to the total number of sampling points in the first power frequency cycle; S2.4: Obtain the disturbance power amplitude, and combine it with the grid rated frequency and the peak value of the frequency change rate to obtain the system total inertia time constant identification value in terms of time.

5. The method for optimizing and tuning parameters of a self-synchronizing voltage source energy storage controller according to claim 1, characterized in that, Step S3 includes: S3.1: Obtain the maximum allowable rate of change limit specified in the power grid operation specifications, and determine the basic inertia requirement value required to maintain system frequency security in combination with the rated frequency of the power grid; S3.2: Based on the ratio of the peak frequency change rate to the maximum allowable frequency change rate limit, the basic inertia requirement value is proportionally amplified to obtain a system inertia requirement value that is adapted to the current severity of the disturbance; S3.3: Calculate the difference between the system inertia requirement value and the system total inertia time constant identification value to obtain the initial inertia shortage; S3.4: Generate a voltage correction factor based on the per-unit value of the grid connection point voltage amplitude, and fuse the initial inertia shortage with the voltage correction factor to obtain the system inertia shortage, wherein the lower the per-unit value of the grid connection point voltage amplitude, the larger the voltage correction factor; S3.5: Perform low-pass filtering on the system inertia shortage and use the filtered system inertia shortage as the interface quantity with the integral feedback controller in step S4.

6. The method for optimizing and tuning parameters of a self-synchronizing voltage source energy storage controller according to claim 1, characterized in that, Step S4 includes: S4.1: Obtain the rated power capacity and default virtual inertia time constant of the energy storage converter, and determine the initial setting value of the integral coefficient based on the rated power capacity; S4.2: Obtain the current output active power of the energy storage converter, combine it with the rated power capacity, obtain the available power margin, and correct the initial setting value of the integral coefficient according to the available power margin to obtain the corrected integral coefficient. S4.3: Determine the output limiting boundary of the integral feedback controller based on the upper and lower limits of the adjustable range of the virtual inertia time constant; S4.4: Using the system inertia shortage as input and the corrected integral coefficient as integral gain, construct an integral operation unit, wherein the output of the integral operation unit is the virtual inertia time correction amount; S4.5: Construct a limiting unit, and connect the output of the integral operation unit to the limiting unit. The limiting unit performs limiting processing on the virtual inertia time correction amount according to the upper limit value and the lower limit value of the output. S4.6: Construct a coefficient adaptive update unit and obtain the ratio of the updated virtual inertia time constant to the current virtual inertia time constant. Adjust the corrected integral coefficients according to the ratio and output the adjusted integral coefficients to the integral operation unit for integral operation in the next control cycle.

7. The method for optimizing and tuning parameters of a self-synchronizing voltage source energy storage controller according to claim 6, characterized in that, The integral feedback controller is composed of an integral coefficient correction unit, an integral calculation unit, a limiting unit, and a coefficient adaptive update unit connected together. The integral coefficient correction unit is connected to the rated power capacity, the default virtual inertia time constant, and the current output active power. The integral calculation unit is connected to the system inertia shortage. The limiting unit outputs the limited energy storage virtual inertia time correction. The coefficient adaptive update unit is connected to the updated virtual inertia time constant and feeds back the adjusted integral coefficient to the integral calculation unit.

8. The method for optimizing and tuning parameters of a self-synchronizing voltage source energy storage controller according to claim 1, characterized in that, Step S5 includes: S5.1: Sum the current virtual inertia time constant with the limited energy storage virtual inertia time correction amount to obtain the updated virtual inertia time constant; S5.2: Compare the updated virtual inertia time constant with the upper and lower limits of the adjustable range of the virtual inertia time constant. If it is determined that the virtual inertia time constant exceeds the upper limit or falls below the lower limit, the upper limit or the lower limit is used as the updated virtual inertia time constant, respectively; otherwise, it remains unchanged. S5.3: Calculate the ratio of the updated virtual inertia time constant to the current virtual inertia time constant, and use it as the droop coefficient synchronous adjustment scaling factor; S5.4: Calculate the voltage support bias factor based on the per-unit value of the voltage amplitude at the grid connection point; S5.5: Determine the updated droop coefficient based on the current droop coefficient, the droop coefficient synchronous adjustment ratio factor, and the voltage support bias factor; S5.6: Combine the updated virtual inertia time constant after the limiting process in S5.2 with the updated droop coefficient to form an optimized control parameter group, and replace the current virtual inertia time constant and current droop coefficient in the self-synchronizing voltage source control loop of the energy storage converter.

9. The method for optimizing and tuning parameters of a self-synchronizing voltage source energy storage controller according to claim 8, characterized in that, Step S5.4 includes: S5.4.1: A dead zone interval is constructed with the rated voltage per unit value as a reference point. When the voltage amplitude per unit value at the grid connection point is within the dead zone interval, the voltage support bias factor is set to 1. The dead zone interval is centered on the rated voltage per unit value. The upper limit is the sum of the rated voltage per unit value and the allowable positive voltage deviation value in the grid operation specification. The lower limit is the difference between the rated voltage per unit value and the allowable negative voltage deviation value in the grid operation specification. The closed interval between the upper limit and the lower limit constitutes the dead zone interval. S5.4.2: When the per-unit value of the grid connection point voltage amplitude is lower than the lower limit of the dead zone, the difference between the per-unit value of the grid connection point voltage amplitude and the lower limit of the dead zone is calculated as the voltage deviation, and the voltage support bias factor is the sum of 1 and the voltage deviation. S5.4.3: When the per-unit value of the grid connection point voltage amplitude is higher than the upper limit of the dead zone interval, the difference between the upper limit of the dead zone interval and the per-unit value of the grid connection point voltage amplitude is calculated as the voltage over-limit amount. The voltage support bias factor is the difference between 1 and the voltage over-limit amount. If the calculation result is negative, the voltage support bias factor is 0.

10. A parameter optimization and tuning system for a self-synchronizing voltage source energy storage controller, characterized in that, The system is used to perform the parameter optimization and tuning method for a self-synchronizing voltage source energy storage controller according to any one of claims 1 to 9, the system comprising: The event detection and data capture module is used to continuously monitor the active power fluctuation amplitude of the energy storage grid connection point. When the active power fluctuation amplitude exceeds the trigger threshold, the current time is recorded as the trigger time, and the frequency change curve of the grid connection point within a preset time period is captured from the trigger time. The total inertia online identification module is used to calculate the system total inertia time constant identification value under the current power grid operating state based on the peak value of the frequency change rate of the frequency change curve at the grid connection point within the first power frequency cycle after the triggering time, combined with the active power fluctuation amplitude. The inertia shortage assessment module is used to calculate the system inertia requirement required to maintain system frequency security based on the current grid rated frequency and the maximum allowable frequency change rate constraint, and to determine the system inertia shortage by combining the system total inertia time constant identification value. The correction calculation and limiting module is used to input the system inertia shortage into the integral feedback controller, output the virtual inertia time correction of the energy storage, and limit the virtual inertia time correction of the energy storage converter with the rated power capacity of the energy storage converter as the limiting condition. The control parameter update module is used to update the virtual inertia time constant of the energy storage according to the virtual inertia time correction after the limit, and to update the droop coefficient synchronously based on the virtual inertia time constant before and after the update, so as to obtain the updated control parameter combination.

Images