A method for suppressing low-frequency oscillation of a photovoltaic energy storage system

Abstract

The present application belongs to the technical field of photovoltaic energy storage grid-connected control, and relates to a low-frequency oscillation suppression method for a photovoltaic energy storage system, which comprises the following steps: collecting reactive power on the photovoltaic side and the energy storage side and the frequency at the grid-connected point, demodulating and separating the reactive power oscillation component and the frequency oscillation component through orthogonal rotation; obtaining the amplitude response ratio and the phase shift of each side through correlation analysis, comparing the slope of the amplitude response ratio with respect to the oscillation frequency to determine the dominant disturbance source, and determining the damping power distribution coefficient; calculating the equivalent reactance of the power grid, generating the damping power command from the damping power distribution coefficient and the oscillation envelope, correcting the time sequence in advance in combination with the phase shift on the side of the dominant disturbance source, and injecting the corresponding side power control outer loop to execute the damping power injection after limiting the amplitude of the remaining capacity of the energy storage and the adjustable power margin of the photovoltaic. The present application realizes the identification of the dominant oscillation source and the differentiated damping distribution, corrects the lag through phase pre-compensation, and avoids the damping internal consumption caused by fixed proportional distribution.

Description

Technical Field

[0001] This invention belongs to the field of photovoltaic energy storage grid-connected control technology, and relates to a method for suppressing low-frequency oscillations in photovoltaic energy storage systems. Background Technology

[0002] During grid-connected operation of photovoltaic energy storage systems, random fluctuations in photovoltaic output, changes in grid impedance, and interactive coupling between converter control loops can easily cause low-frequency oscillations in the active power at the grid connection point within the 0.1Hz to 2.5Hz frequency band, leading to decreased system stability and malfunctions of protection devices.

[0003] In the existing technology, the methods for suppressing this type of oscillation mainly include two categories: First, a virtual inertia damping element is introduced into the active power control loop of the grid-connected converter, and a damping power command is generated through a fixed proportional gain of the frequency deviation; Second, the power system stabilizer architecture is used to perform Fourier analysis on the grid connection point frequency to extract the oscillation mode, and then a damping signal is generated through a phase compensation network with fixed parameters.

[0004] However, the existing methods mentioned above have the following drawbacks in practical applications: 1. Lack of ability to identify the dominant source of oscillation energy: When the DC voltage loop disturbance on the photovoltaic side and the power loop disturbance on the energy storage side coexist and are coupled with each other, the existing methods cannot identify the oscillation driving source and the driven side, resulting in a lack of basis for selecting the action point of damping injection, and the suppression effect depends on trial and error.

[0005] 2. Phase compensation has an inherent lag. Existing methods rely on Fourier transform to obtain oscillation phase information, and the data window length is at least one oscillation period, causing the compensated phase to lag behind the actual oscillation phase. When the oscillation frequency deviates due to operating conditions, the lag causes the compensated damping torque to be out of phase with the oscillation component, and in severe cases, it will inject negative damping and aggravate the oscillation.

[0006] 3. Existing multi-machine collaborative allocation strategies are crude: When photovoltaic and energy storage participate in suppression at the same time, existing methods use a fixed ratio or equal allocation method, without considering the difference in control bandwidth and operating boundary constraints on both sides. This results in an additional phase difference between the outputs on both sides, weakening the overall damping effect. Summary of the Invention

[0007] In view of this, in order to solve the problems mentioned in the background technology, a method for suppressing low-frequency oscillations in photovoltaic energy storage systems is proposed.

[0008] The objective of this invention can be achieved through the following technical solution: This invention provides a method for suppressing low-frequency oscillations in a photovoltaic energy storage system, comprising: synchronously acquiring the reactive power output from the photovoltaic side and the energy storage side and the grid connection point frequency, and separating the reactive power oscillation component and the frequency oscillation component through orthogonal rotation demodulation.

[0009] Correlation analysis is performed on the reactive power oscillation component and the frequency oscillation component to obtain the first amplitude response ratio and the first phase shift on the photovoltaic side, and the second amplitude response ratio and the second phase shift on the energy storage side.

[0010] By comparing the slopes of the first and second amplitude response ratios as a function of oscillation frequency, the dominant disturbance source is identified and a damping weight allocation logic is established to determine the damping power allocation coefficients for the photovoltaic and energy storage sides.

[0011] The equivalent reactance of the grid is calculated by using the transient relationship between voltage and power at the grid connection point. The damping power command is generated by the damping power allocation coefficient and the oscillation envelope calculation. The phase pre-compensation amount is generated by combining the phase offset on the side of the dominant disturbance source to pre-correct the timing of the damping power command.

[0012] The corrected damping power command is sent to the corresponding power control outer loop. After the remaining energy storage capacity and the adjustable power margin of the photovoltaic DC side limit the command of each side, the damping power injection is performed.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) The present invention calculates the amplitude response ratio and phase shift between reactive power disturbance and frequency disturbance on the photovoltaic side and the energy storage side, and compares the slope of the amplitude response ratio on both sides with the change of oscillation frequency to determine the dominant disturbance source type. It can identify the driving source and the driven side from the mutually coupled oscillation circuit, so that the selection of the damping injection point has clear criteria and avoids relying on trial and error to determine the damping allocation object.

[0014] (2) This invention utilizes the relationship between the voltage amplitude at the grid connection point and the transient change in reactive power to calculate the equivalent reactance of the power grid online. Combined with the phase shift on the side of the dominant disturbance source, a phase pre-compensation amount is generated to perform timing pre-correction on the damping power command, compensate for the phase lag introduced by the signal processing link and the power grid impedance, so that the damping torque and the oscillation component remain in phase, and avoid the injection of negative damping due to phase mismatch.

[0015] (3) Based on the determination result of the dominant disturbance source, the present invention establishes the damping weight allocation logic, reduces the damping intervention weight on the dominant disturbance source side and increases the damping intervention weight on the auxiliary side, generates the damping power allocation coefficient on both sides, and combines the remaining energy storage capacity and the adjustable power margin of the photovoltaic DC side to limit the output of the command, which helps to achieve differentiated coordination of the damping output on both sides, avoids the internal phase difference and damping internal loss caused by fixed ratio allocation, and ensures that the damping command can be executed within the equipment operating boundary. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating the implementation steps of the method of the present invention;

[0018] Figure 2 This is a schematic diagram illustrating the logic for determining the dominant disturbance source in this invention;

[0019] Figure 3 This is a schematic diagram of the damping weight allocation logic of the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Please see Figure 1 As shown, the present invention provides a method for suppressing low-frequency oscillations in a photovoltaic energy storage system, comprising: S1. synchronously acquiring the reactive power output from the photovoltaic side and the energy storage side and the grid connection point frequency, and separating the reactive power oscillation component and the frequency oscillation component through orthogonal rotation demodulation.

[0022] In this embodiment, separating the reactive power oscillation component and the frequency oscillation component includes: constructing a synchronous rotating orthogonal coordinate system with the grid connection point frequency as the rotational angular velocity; reading the grid connection point frequency output by the phase-locked loop at each sampling moment, multiplying the grid connection point frequency by twice pi as the current rotational angular velocity, multiplying the current rotational angular velocity by a fixed sampling interval, and adding the product to the phase angle of the previous sampling moment to obtain the phase angle of the current sampling moment; the phase angle represents the real-time rotation angle of the synchronous rotating orthogonal coordinate system relative to the stationary coordinate system, and its numerical range is [missing information]. When the phase angle exceeds at any sampling time Automatic subtraction is required. .

[0023] The reactive power is constructed using an orthogonal signal. Specifically, a circular buffer is established to store the reactive power values ​​at several consecutive sampling times. The buffer length is taken as a fraction of the number of sampling points corresponding to the rated frequency period of the grid connection point. Round up.

[0024] Read the latest reactive power value stored in the current buffer as the first reactive power component; read the reactive power value at the moment corresponding to the time when it is shifted forward by half the buffer length as the second reactive power component. The first reactive power component and the second reactive power component are 90 degrees out of phase, forming an orthogonal signal pair.

[0025] When the number of sampling points stored in the buffer is less than half the length of the buffer, the slope of the reactive power change per unit time is obtained by dividing the difference of the reactive power values ​​of the two most recent sampling times by the fixed sampling interval. For the sampling times to be filled, starting from the reactive power value of the most recent sampling time, the product of the slope of change and the fixed sampling interval is accumulated one by one to fill the unfilled sampling points in the buffer until the data requirement of half the length of the buffer is met.

[0026] Based on the phase angle at the current sampling time, the reactive cross-axis component is calculated according to the following rotational transformation formula: .

[0027] in, This is the reactive quadrature-axis component, used to reflect the oscillating component in reactive power that deviates from the fundamental frequency of the power grid. The phase angle at the current sampling time. The first reactive component, This is the second reactive component.

[0028] Project the first reactive component onto the... The relevant coordinate axes are marked with a negative sign to cancel or separate the fundamental reactive component.

[0029] Project the second reactive component onto the... On the relevant coordinate axes, reactive oscillation information orthogonal to the fundamental wave is extracted.

[0030] The physical meaning of the entire formula lies in realizing the rotational coordinate transformation, and demodulating the reactive oscillation component reflecting the frequency deviation in the actual physical coordinate system for the first reactive component and the second reactive component.

[0031] Frequency band constraint extraction is performed on the reactive quadrature axis component to filter out the fundamental and high-frequency components, and retain the disturbance signal within the preset low-frequency oscillation band as the reactive power oscillation component. Specifically, the reactive quadrature axis component is sent to a bandpass filter, where the lower cutoff frequency of the passband is the preset lower limit of the low-frequency oscillation band (0.1Hz) and the upper cutoff frequency is the preset upper limit of the low-frequency oscillation band (2.5Hz). The output after filtering is used as the reactive power oscillation component.

[0032] The deviation between the grid connection point frequency and its rated value is defined as the frequency deviation. Frequency band constraint extraction is performed on the frequency deviation to obtain the frequency oscillation component.

[0033] S2. Perform correlation analysis on the reactive power oscillation component and the frequency oscillation component to obtain the first amplitude response ratio and the first phase shift on the photovoltaic side, and the second amplitude response ratio and the second phase shift on the energy storage side.

[0034] In this embodiment, the correlation analysis between the reactive power oscillation component and the frequency oscillation component includes: applying a sliding time window to the reactive power oscillation component and the frequency oscillation component, using a Hanning window as the window function; dividing the sampling frequency by a preset lower limit frequency of the low-frequency oscillation band to obtain the number of sampling points per cycle; rounding the window length up to the nearest power of two, taking twice the number of sampling points per cycle; and using a sliding step size equal to a fraction of the window length. .

[0035] Perform a Fast Fourier Transform on the windowed data sequence within each window to obtain the spectral amplitude distribution of the corresponding window. Arrange the spectra of each window along the time axis to form the time spectrum.

[0036] In the time spectrum, the root mean square value of the amplitude along the time axis at each frequency point is taken as the global energy.

[0037] The global energy at each frequency point is compared with a preset energy threshold one by one. The frequency band corresponding to the frequency point where the global energy is greater than the preset energy threshold is identified as a dominant oscillation mode. If multiple dominant oscillation modes are identified, the frequency band corresponding to the dominant oscillation mode with the largest global energy is selected as the dominant oscillation frequency band; if there is only one dominant oscillation mode, its corresponding frequency band is directly selected as the dominant oscillation frequency band. The dominant oscillation frequency band is represented by the center frequency and bandwidth.

[0038] The preset energy threshold is determined as follows: under the operating condition where the frequency deviation at the grid connection point is continuously below 0.05Hz and the reactive power fluctuation does not exceed 1% of the rated capacity, reactive power oscillation component data is continuously collected for at least 1 hour. After performing STFT analysis on the data, the maximum value of the global energy at all frequency points is multiplied by the expansion factor to obtain the preset energy threshold; wherein the expansion factor is set to 3.

[0039] Within the dominant oscillation frequency band, cross-spectral analysis is performed on the reactive power oscillation component and the frequency oscillation component on the photovoltaic side. The calculation process of the cross-spectral analysis is as follows: take the fast Fourier transform results of the reactive power oscillation component and the frequency oscillation component on the photovoltaic side within the current sliding window, multiply the spectral value of the reactive power oscillation component on the photovoltaic side with the conjugate of the spectral value of the frequency oscillation component at each frequency point to obtain the cross-spectral density on the photovoltaic side; at the same time, calculate the product of the spectral value of the frequency oscillation component and its own conjugate to obtain the self-spectral density of the frequency oscillation component.

[0040] Within the dominant oscillation frequency band, the frequency point with the largest amplitude of the photovoltaic-side cross-spectral density is determined as the spectral peak frequency. The ratio of the amplitude of the photovoltaic-side cross-spectral density at the spectral peak frequency to the amplitude of the frequency oscillation component from the spectral density is used as the first amplitude response ratio. The phase angle of the photovoltaic-side cross-spectral density at the spectral peak frequency is used as the first phase shift. The phase angle is obtained by taking the arctangent of the ratio of the imaginary part to the real part of the cross-spectral density, and quadrant correction is performed based on the signs of the imaginary and real parts.

[0041] By replacing the reactive power oscillation component on the photovoltaic side with the reactive power oscillation component on the energy storage side, the second amplitude response ratio and the second phase shift on the energy storage side are obtained using the exact same process.

[0042] S3. Compare the slopes of the first amplitude response ratio and the second amplitude response ratio as a function of oscillation frequency, determine the dominant disturbance source, establish damping weight allocation logic, and determine the damping power allocation coefficient between the photovoltaic side and the energy storage side.

[0043] Reference Figure 2 As shown, in this embodiment, determining the dominant disturbance source includes: performing point-by-point first-order difference operations on the first amplitude response ratio sequence and the second amplitude response ratio sequence on the discrete frequency sequence corresponding to the dominant oscillation mode, respectively, to obtain the first slope sequence and the second slope sequence; wherein, the first-order difference operation process is: for the first slope sequence in the sequence... For each frequency point, take the amplitude response ratio corresponding to the next frequency point, subtract the amplitude response ratio corresponding to the current frequency point, and divide by the frequency resolution to obtain the i-th frequency point. First-order difference values ​​at each frequency point.

[0044] Calculate the difference between the absolute value of the first slope and the absolute value of the second slope at each frequency point within the dominant oscillation frequency band, and construct a slope difference sequence.

[0045] Set a preset number of sliding windows on the slope difference sequence, and count the number of elements with positive and negative difference signs in each sliding window.

[0046] When the number of elements with positive values ​​in each sliding window is greater than the number of elements with negative values, the dominant disturbance source is determined to be the photovoltaic side; when the number of elements with negative values ​​in each sliding window is greater than the number of elements with positive values, the dominant disturbance source is determined to be the energy storage side; when the average value of all elements in each sliding window is less than the preset tolerance threshold, the dominant disturbance source is determined to be the external power grid.

[0047] It should be noted that the preset tolerance threshold is determined as follows: Events where the photovoltaic or energy storage side was successfully identified as the dominant disturbance source are retrieved. The maximum absolute value of the slope difference for each event within each sliding window is recorded. The median value of all records is obtained and multiplied by an attenuation coefficient to obtain the preset tolerance threshold. The attenuation coefficient is between 0.2 and 0.5, determined by the acceptable external disturbance sensitivity on-site. If there are no historical records for the first operation, the absolute values ​​of each element in the slope difference sequence within the dominant oscillation frequency band are arranged in ascending order, and the value located at the lower quartile is taken as the preset tolerance threshold.

[0048] If none of the sliding windows meet the conditions that the number of elements is positive and greater than or negative, or that the number of elements is negative and greater than or positive, and the average value of the elements is not less than the preset tolerance threshold, then the determination result of the dominant disturbance source of the previous determination period remains unchanged.

[0049] Reference Figure 3 As shown, in this embodiment, the establishment of damping weight allocation logic includes: calculating the arithmetic mean of the slope difference sequence within the dominant oscillation frequency band over a consecutive preset number of sliding windows, as an index of damping imbalance.

[0050] The basic offset is obtained by normalizing the damping imbalance index. The specific normalization process is as follows: at each frequency point within the dominant oscillation frequency band, the difference between the first slope sequence and the second slope sequence is calculated separately as the original slope difference value, and its absolute value is taken; the absolute values ​​of the original slope differences corresponding to all frequency points are arranged in descending order, and the value located at the upper quartile is taken as the reference value; the current damping imbalance index is compared with the reference value, where if the reference value is less than a preset positive number (value is taken as...), the offset is determined by... If the base offset is 0, then the base offset will be set to 0.

[0051] Obtain the current adjustable power limit of the photovoltaic side and the current adjustable power limit of the energy storage side, and calculate the proportion of the adjustable power limit of the photovoltaic side to the total adjustable power limit, which is used as the capacity weight of the photovoltaic side.

[0052] Based on different dominant disturbance sources, the basic offset and the photovoltaic side capacity weight are cross-calculated: when the dominant disturbance source is the photovoltaic side, the energy storage side damping power allocation coefficient is taken as the sum of the photovoltaic side capacity weight and the basic offset.

[0053] When the dominant disturbance source is the energy storage side, the damping power allocation coefficient of the energy storage side is taken as the difference between the photovoltaic side capacity weight and the foundation offset.

[0054] When the dominant disturbance source is the external power grid, the damping power distribution coefficient of the energy storage side is taken as the preset equilibrium base point value. The preset equilibrium base point value is determined as follows: read the rated capacity of the photovoltaic inverter and the rated capacity of the energy storage converter from the equipment nameplate parameters, take the rated capacity of the energy storage converter as the numerator, and the sum of the rated capacity of the photovoltaic inverter and the rated capacity of the energy storage converter as the denominator. The resulting quotient is taken as the preset equilibrium base point value.

[0055] The photovoltaic-side damping power allocation coefficient is obtained by subtracting the energy storage-side damping power allocation coefficient from 1.

[0056] It should be noted that the calculated energy storage-side damping power allocation coefficient is... Cut-off limiting; when cut-off occurs, the photovoltaic-side damping power distribution coefficient is adjusted accordingly to keep the sum of the coefficients on both sides at 1.

[0057] S4. Calculate the equivalent reactance of the grid by using the transient relationship between the voltage and power at the grid connection point. Generate a damping power command by calculating the damping power distribution coefficient and the oscillation envelope. Generate a phase pre-compensation amount by combining the phase offset on the side of the dominant disturbance source. Perform timing correction on the damping power command.

[0058] In this embodiment, the step of using the transient variation relationship between grid connection point voltage and power to calculate the grid equivalent reactance includes: continuously monitoring the voltage amplitude sequence at the grid connection point, and marking the current sampling point as the disturbance start point when the voltage amplitude change between adjacent sampling points exceeds the peak value of the voltage amplitude fluctuation in the previous sampling interval.

[0059] Starting from the disturbance initiation point, the ratio of voltage amplitude change to active power change at each sampling point is continuously calculated. When the ratio first shows a sign reversal, and the subsequent sampling points maintain a sign reversal for a preset number of consecutive sampling points, the sampling point where the sign first reverses is identified as the disturbance recovery point.

[0060] Within the interval from the disturbance initiation point to the disturbance recovery point, calculate the absolute value of the ratio of voltage amplitude change to active power change at each sampling point, locate the sampling point with the largest absolute value, and use the ratio of voltage amplitude change to active power change corresponding to it as the equivalent reactance of the power grid.

[0061] It should be noted that the voltage amplitude acquisition process is as follows: the three-phase voltage at the grid connection point output by the phase-locked loop is transformed by rotating coordinates to obtain the direct-axis voltage component and the quadrature-axis voltage component; the sum of squares of the direct-axis voltage component and the quadrature-axis voltage component is calculated, and then the square root of the sum of squares is taken to obtain the voltage amplitude.

[0062] The previous sampling interval is a continuous sampling interval preceding the current sampling point, with a duration equal to twice the dominant oscillation period. The peak-to-peak value of the voltage amplitude fluctuation refers to the difference between the maximum and minimum voltage amplitude values ​​within the previous sampling interval.

[0063] In this embodiment, the phase pre-compensation amount is generated through the following process: obtaining the phase offset value on one side of the dominant disturbance source, wherein the photovoltaic side corresponds to the first phase offset and the energy storage side corresponds to the second phase offset.

[0064] Multiply the rated voltage at the grid connection point by itself to obtain the square of the rated voltage; calculate the quotient of the grid equivalent reactance and the square of the rated voltage, and multiply the quotient by the current dominant oscillation angular frequency to obtain the inherent phase shift angle caused by the grid impedance.

[0065] The total hysteresis phase angle is obtained by adding the phase offset value to the inherent phase shift angle.

[0066] Perform a tangent operation on the total lag phase angle, and use the ratio of the obtained tangent value to the dominant oscillation angular frequency as the phase pre-compensation amount, the sign of which is opposite to that of the total lag phase angle.

[0067] It should be noted that the physical meaning of the phase precompensation amount is: to reduce the total lag phase angle. Convert to equivalent time-domain lead time ,satisfy The relationship, among which As the dominant oscillation angular frequency, therefore This lead time reflects the amount of timing advance required for the damped envelope signal to achieve phase alignment.

[0068] The phase pre-compensation amount has a saturation limit: the upper limit is half of the dominant oscillation period, and the lower limit is a negative value of half of the dominant oscillation period; when the calculated result of the phase pre-compensation amount exceeds the saturation limit, the saturation value in the corresponding direction is taken.

[0069] In this embodiment, the timing correction of the damping power command includes: rounding down the ratio of the phase pre-compensation amount to the preset sampling period to obtain the number of advance sampling points.

[0070] After taking the absolute value of the reactive power oscillation component, a sliding window mean filter is applied. Specifically, after taking the absolute value of the reactive power oscillation component, the rectified oscillation signal is obtained. The number of sampling points corresponding to half of the dominant oscillation period is taken as the window length, and a first-in-first-out buffer with a length equal to the number of sampling points of the window length is established. At each sampling moment, the current rectified oscillation signal value is stored at the end of the buffer, while the earliest stored value at the beginning of the buffer is removed, so that the buffer always retains the rectified oscillation signal value of the most recent window length sampling moments. The arithmetic mean of all values ​​in the buffer is calculated, and the average value is output as the oscillation envelope signal of the current sampling moment. The storage, removal, and averaging operations are repeated at the next sampling moment, so as to realize that the window slides point by point along the time axis and continuously outputs the envelope signal.

[0071] A timing offset is applied to the oscillating envelope signal. Based on the current sampling time, the oscillating envelope values ​​corresponding to the number of leading sampling points after the current time are advanced to the current time and output to form a leading envelope signal.

[0072] Obtain the photovoltaic-side damping power allocation coefficient and the energy storage-side damping power allocation coefficient at the current moment, and multiply them by the leading envelope signal to obtain the photovoltaic-side damping fundamental component and the energy storage-side damping fundamental component.

[0073] The product of the frequency deviation and the preset gain coefficient is determined as the frequency recovery component. The preset gain coefficient is determined as follows: the maximum allowable frequency deviation of the power grid is read from the grid connection standard, with a typical value of 0.5Hz. The sum of the total rated power of the photovoltaic side and the energy storage side is divided by the maximum frequency deviation to obtain the preset gain coefficient.

[0074] The photovoltaic-side damping fundamental component and the frequency recovery component are added together to obtain the photovoltaic-side damping power command; the energy storage-side damping fundamental component and the frequency recovery component are added together to obtain the energy storage-side damping power command.

[0075] It should be noted that the total lag phase angle is formed by the superposition of the phase shift of the dominant disturbance source and the inherent phase shift caused by the grid impedance, representing the physical delay from oscillation sensing to the actual effectiveness of the damping command. The sign of the phase pre-compensation amount is opposite to that of the total lag phase angle. When the total lag phase angle is positive, the pre-compensation amount is negative, corresponding to a timing lead correction, meaning the damping power command needs to be issued before the oscillation peak. When the total lag phase angle is negative due to special operating conditions such as grid capacitive behavior, the pre-compensation amount is positive, corresponding to a timing lag correction. This embodiment uses a typical operating condition with a positive total lag phase angle as an example to illustrate the execution process of lead compensation.

[0076] S5. The corrected damping power command is sent to the corresponding power control outer loop. After the remaining energy storage capacity and the adjustable power margin of the photovoltaic DC side limit the command of each side, the damping power injection is performed.

[0077] In this embodiment, limiting the photovoltaic-side damping power command includes multiplying the difference between the current voltage value of the photovoltaic DC side and the voltage at the maximum power point by a preset conversion coefficient to obtain the photovoltaic forward adjustable power margin.

[0078] The preset conversion coefficient is used to convert the DC side voltage deviation into an adjustable AC power margin. The determination method is as follows: during the factory testing phase of the photovoltaic inverter, multiple typical operating conditions are set in sequence, and each operating condition is composed of a preset irradiance and a preset ambient temperature.

[0079] Under each operating condition, the photovoltaic simulation power supply is adjusted so that the DC side voltage of the inverter gradually approaches the theoretical maximum power point voltage of the current operating condition. The DC side voltage and AC output power are continuously collected. When the peak-to-peak value of the DC side voltage is lower than the preset voltage fluctuation threshold within three adjacent sampling periods, the current DC side voltage is recorded as the maximum power point voltage of the corresponding operating condition, and the current AC output power is recorded as the reference power.

[0080] Keeping the operating conditions unchanged, within the adjustable range of the DC side voltage, with the maximum power point voltage as the reference, select multiple voltage offset points at equal intervals in the directions above and below the maximum power point voltage.

[0081] For each voltage offset point, the voltage setpoint of the DC converter is set as the target voltage corresponding to the offset point. When the absolute value of the difference between the DC side voltage sample value and the target voltage is less than the preset voltage deviation tolerance in multiple consecutive control cycles, the current AC output power is recorded, the difference between the AC output power at the offset point and the reference power is calculated as the power change, and the difference between the target voltage and the voltage at the maximum power point is calculated as the voltage deviation.

[0082] Calculate the ratio of power change to voltage deviation to obtain the single-point conversion coefficient for each voltage offset point.

[0083] The average value of the single-point conversion coefficients for all offset points under the same working condition is used to obtain the working condition conversion coefficient.

[0084] The average value of the condition conversion coefficients for all typical working conditions is taken to obtain the preset conversion coefficients.

[0085] Multiply the difference between the maximum power point voltage and the current DC voltage of the photovoltaic system by a preset conversion factor to obtain the photovoltaic negative adjustable power margin.

[0086] When the photovoltaic-side damping power command is greater than the photovoltaic positive adjustable power margin, the photovoltaic positive adjustable power margin is used; otherwise, the original value is used. Then, the result is compared with the negative value of the photovoltaic negative adjustable power margin for boundary judgment. When the result is less than the negative value of the photovoltaic negative adjustable power margin, the negative value of the photovoltaic negative adjustable power margin is used; otherwise, the original value is used, thus completing the photovoltaic-side command limiting.

[0087] In this embodiment, limiting the damping power command on the energy storage side includes: multiplying the installed energy storage capacity by the current state of charge value to obtain the remaining rechargeable capacity, and multiplying the difference between the fully charged state value and the current state of charge value by the installed energy storage capacity to obtain the remaining dischargeable capacity.

[0088] When the energy storage side damping power command is greater than the remaining dischargeable capacity, the remaining dischargeable capacity is used; otherwise, the original value is used. Then, the result is compared with the negative value of the remaining rechargeable capacity for boundary judgment. When the result is lower than the negative value of the remaining rechargeable capacity, the negative value of the remaining rechargeable capacity is used; otherwise, the original value is used, thus completing the energy storage side command limiting.

[0089] Calculate the absolute value of the difference between the energy storage side damping power command after the limit is set and the value of the previous control cycle. When the absolute value of the difference is greater than the preset change limit, the current command is corrected to the algebraic sum of the value of the previous cycle and the preset change limit. The sign of the corrected command is consistent with that of the command before the correction.

[0090] It should be noted that the preset variation limit is determined as follows: the maximum allowable charge and discharge rate of the energy storage battery is read from the battery's factory parameters, the sampling period is converted to hourly units, and then accumulated with the maximum charge and discharge rate and the rated power of the energy storage converter to obtain the preset variation limit. Its physical meaning is the maximum allowable power change of the battery within one sampling period.

[0091] When continuous If the damping power command triggers limiting within each control cycle, it indicates that the current adjustable power margin on both sides is insufficient to meet the damping power requirement. The communication sub-board generates a limiting alarm signal and simultaneously activates a degradation strategy: the damping power command is gradually reduced according to a preset attenuation factor, and the passband of the bandpass filter is narrowed from a preset low-frequency oscillation band to the dominant oscillation band, prioritizing damping injection into the dominant oscillation band to prevent complete loss of damping due to full limiting; among which... The control period is determined by the center frequency of the current dominant oscillation mode and the control period of the power control outer loop. Specifically, the product of the center frequency of the dominant oscillation band and the control period of the power control outer loop is differentiated to obtain the number of control periods corresponding to a single dominant oscillation period. This number of control periods is then rounded up and multiplied by 2, and the result is used as the control period. The value of .

[0092] The preset decay factor acquisition process is as follows: backtrack to the nearest For each control cycle, the limiting records are used to extract the pre-limiting damped power command value and the actual injected value after limiting on the photovoltaic side, and the pre-limiting damped power command value and the actual injected value after limiting on the energy storage side. For the photovoltaic side, the ratio of the actual injected value after limiting to the pre-limiting damped power command value is calculated cycle by cycle to obtain... The single-cycle injection ratio on the photovoltaic side is taken as follows: The arithmetic mean of the proportions is used as the photovoltaic-side degradation factor. The same method is used to calculate the degradation factor for the energy storage side. The arithmetic mean of the single-cycle injection ratio of each energy storage side is used as the energy storage side attenuation factor. The physical meaning of the attenuation factor is the average ratio of the actual injectable damping power to the damping demand power under the current operating conditions, and the value range is greater than 0 and less than or equal to 1. When the calculated attenuation factor is lower than the preset attenuation lower limit, the preset attenuation lower limit is taken as the attenuation factor of that side, and the preset attenuation lower limit is 0.3.

[0093] Since the downgrade strategy was initiated, every time... Each control cycle reassesses the amplitude limiting status; if continuous Neither the photovoltaic-side damping power command nor the energy storage-side damping power command triggered the limiting during the control cycle, thus exiting the degradation strategy.

[0094] It should also be noted that this invention has two limitations on the use of preset quantities. First, a preset number of sliding windows are set on the slope difference sequence for multi-point joint statistics of the slope difference sign distribution in the determination of the dominant disturbance source. The typical value of the preset number of sliding windows is 3 to 7, preferably 5. Second, in the process of confirming the disturbance recovery point, a preset number of sampling points are required to maintain the sign reversal after the ratio first shows a reversal. This is used to exclude false reversals caused by noise or transient ripple jittering near zero. The preset number of sampling points is typically the rounded-up value of the number of sampling points corresponding to half a dominant oscillation cycle. For example, when the dominant oscillation cycle is 2 seconds and the sampling cycle is 0.02 seconds, the preset number is 50 sampling points. Both preset quantities serve the same design principle: replacing single-point instantaneous decision with continuous multi-point joint decision, improving robustness against transient noise and local spikes and reducing the probability of misjudgment by confirming the consistency of samples over a certain time span or frequency span.

[0095] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, and all such modifications and additions should fall within the protection scope of the present invention.

Claims

1. A method for suppressing low-frequency oscillations in a photovoltaic energy storage system, characterized in that, include: The reactive power output from the photovoltaic side and the energy storage side and the grid connection point frequency are collected simultaneously. After orthogonal rotation demodulation, the reactive power oscillation component and the frequency oscillation component are separated. Correlation analysis is performed on the reactive power oscillation component and the frequency oscillation component to obtain the first amplitude response ratio and the first phase shift on the photovoltaic side, and the second amplitude response ratio and the second phase shift on the energy storage side. By comparing the slopes of the first and second amplitude response ratios as a function of oscillation frequency, the dominant disturbance source is identified and a damping weight allocation logic is established to determine the damping power allocation coefficients for the photovoltaic side and the energy storage side. The equivalent reactance of the power grid is calculated by using the transient relationship between voltage and power at the grid connection point. The damping power command is generated by the damping power allocation coefficient and the oscillation envelope calculation. The phase pre-compensation amount is generated by combining the phase offset on the side of the dominant disturbance source to pre-correct the timing of the damping power command. The corrected damping power command is sent to the corresponding power control outer loop. After the remaining energy storage capacity and the adjustable power margin of the photovoltaic DC side limit the command of each side, the damping power injection is performed.

2. The method for suppressing low-frequency oscillations in a photovoltaic energy storage system according to claim 1, characterized in that, The separation of reactive power oscillation component and frequency oscillation component includes: A synchronous rotating orthogonal coordinate system is constructed using the grid connection point frequency as the rotation angular velocity; Orthogonal signals are constructed from reactive power to generate a first reactive component and a second reactive component that are orthogonal to each other. The first reactive component and the second reactive component are then transformed to a synchronous rotating orthogonal coordinate system to obtain the reactive cross-axis component. Frequency band constraint extraction is performed on the reactive quadrature axis component to filter out the fundamental and high frequency components, and retain the disturbance signal in the preset low frequency oscillation band as the reactive power oscillation component. The deviation between the grid connection point frequency and its rated value is defined as the frequency deviation. Frequency band constraint extraction is performed on the frequency deviation to obtain the frequency oscillation component.

3. The method for suppressing low-frequency oscillations in a photovoltaic energy storage system according to claim 1, characterized in that, The correlation analysis between the reactive power oscillation component and the frequency oscillation component includes: Perform a sliding window short-time Fourier transform on the reactive power oscillation component and the frequency oscillation component to obtain their respective time spectra, and identify at least one dominant oscillation mode whose energy exceeds a preset energy threshold. When multiple dominant oscillation modes exist, the frequency band corresponding to the dominant oscillation mode with the highest energy is selected as the dominant oscillation frequency band; when only one dominant oscillation mode exists, its corresponding frequency band is selected as the dominant oscillation frequency band. Within the dominant oscillation frequency band, cross-spectral analysis is performed on the reactive power oscillation component and the frequency oscillation component on the photovoltaic side. The amplitude ratio and phase value at the spectral peak are extracted from the cross-spectral results and used as the first amplitude response ratio and the first phase shift, respectively. By comparing the calculation process on the photovoltaic side, the second amplitude response ratio and the second phase offset on the energy storage side are obtained simultaneously.

4. The method for suppressing low-frequency oscillations in a photovoltaic energy storage system according to claim 3, characterized in that, The determination of the dominant disturbance source includes: On the discrete frequency sequence corresponding to the dominant oscillation mode, point-by-point first-order difference operations are performed on the first amplitude response ratio sequence and the second amplitude response ratio sequence to obtain the first slope sequence and the second slope sequence. Calculate the difference between the absolute value of the first slope and the absolute value of the second slope at each frequency point within the dominant oscillation frequency band, and construct a slope difference sequence. Set a preset number of sliding windows on the slope difference sequence, and count the number of elements with positive and negative difference signs in each sliding window; When the number of elements with positive values ​​in each sliding window is greater than the number of elements with negative values, the dominant disturbance source is determined to be the photovoltaic side; when the number of elements with negative values ​​in each sliding window is greater than the number of elements with positive values, the dominant disturbance source is determined to be the energy storage side; when the average value of all elements in each sliding window is less than the preset tolerance threshold, the dominant disturbance source is determined to be the external power grid.

5. The method for suppressing low-frequency oscillations in a photovoltaic energy storage system according to claim 4, characterized in that, The logic for establishing damping weight allocation includes: Calculate the arithmetic mean of the slope difference sequence within the dominant oscillation frequency band over a consecutive preset number of sliding windows, and use it as an index of damping imbalance. The basic offset is obtained after normalizing the damping imbalance index. Obtain the current adjustable power limit of the photovoltaic side and the current adjustable power limit of the energy storage side, and calculate the proportion of the adjustable power limit of the photovoltaic side to the total adjustable power limit as the photovoltaic side capacity weight; Based on different dominant disturbance sources, the basic offset and the photovoltaic capacity weight are cross-calculated: When the dominant disturbance source is the photovoltaic side, the damping power allocation coefficient of the energy storage side is taken as the sum of the photovoltaic side capacity weight and the foundation offset; When the dominant disturbance source is the energy storage side, the damping power allocation coefficient of the energy storage side is taken as the difference between the photovoltaic side capacity weight and the foundation offset. When the dominant disturbance source is the external power grid, the damping power allocation coefficient of the energy storage side is taken as the preset equilibrium base point value; The photovoltaic-side damping power allocation coefficient is obtained by subtracting the energy storage-side damping power allocation coefficient from 1.

6. The method for suppressing low-frequency oscillations in a photovoltaic energy storage system according to claim 1, characterized in that, The calculation of the grid equivalent reactance using the transient relationship between grid connection point voltage and power includes: The voltage amplitude sequence at the grid connection point is continuously monitored. When the change in voltage amplitude between adjacent sampling points exceeds the peak value of voltage amplitude fluctuation in the previous sampling interval, the current sampling point is marked as the disturbance start point. Starting from the disturbance initiation point, the ratio of the change in voltage amplitude to the change in active power at each sampling point is continuously calculated. When the ratio first shows a sign reversal, and the subsequent sampling points all maintain a sign reversal for a preset number of consecutive sampling points, the sampling point where the sign first reverses is identified as the disturbance recovery point. Within the interval from the disturbance initiation point to the disturbance recovery point, calculate the absolute value of the ratio of voltage amplitude change to active power change at each sampling point, locate the sampling point with the largest absolute value, and use the ratio of voltage amplitude change to active power change corresponding to it as the equivalent reactance of the power grid.

7. The method for suppressing low-frequency oscillations in a photovoltaic energy storage system according to claim 1, characterized in that, The phase pre-compensation amount is generated through the following process: Obtain the phase shift value on one side of the dominant disturbance source, where the photovoltaic side corresponds to the first phase shift and the energy storage side corresponds to the second phase shift; Multiply the rated voltage at the grid connection point by itself to obtain the square of the rated voltage; calculate the quotient of the grid equivalent reactance and the square of the rated voltage, and multiply the quotient by the current dominant oscillation angular frequency to obtain the inherent phase shift angle caused by the grid impedance; Add the phase offset value to the inherent phase shift angle to obtain the total hysteresis phase angle; Perform a tangent operation on the total lag phase angle, and use the ratio of the obtained tangent value to the dominant oscillation angular frequency as the phase pre-compensation amount, the sign of which is opposite to that of the total lag phase angle.

8. The method for suppressing low-frequency oscillations in a photovoltaic energy storage system according to claim 2, characterized in that, The timing pre-correction of the damping power command includes: The ratio of the phase pre-compensation amount to the preset sampling period is rounded down to obtain the number of leading sampling points; The absolute value of the reactive power oscillation component is taken and then subjected to sliding window mean filtering to obtain the oscillation envelope signal; A timing offset is applied to the oscillating envelope signal. Based on the current sampling time, the oscillating envelope values ​​corresponding to the number of leading sampling points after the current time are advanced to the current time and output to form a leading envelope signal. Obtain the photovoltaic-side damping power allocation coefficient and the energy storage-side damping power allocation coefficient at the current moment, and multiply them by the leading envelope signal respectively to obtain the photovoltaic-side damping basic component and the energy storage-side damping basic component. The product of the frequency deviation and the preset gain coefficient is determined as the frequency recovery component; The photovoltaic-side damping fundamental component and the frequency recovery component are added together to obtain the photovoltaic-side damping power command; the energy storage-side damping fundamental component and the frequency recovery component are added together to obtain the energy storage-side damping power command.

9. The method for suppressing low-frequency oscillations in a photovoltaic energy storage system according to claim 8, characterized in that, The photovoltaic-side damping power command limit includes: Multiply the difference between the current DC voltage value of the photovoltaic system and the voltage at the maximum power point by a preset conversion factor to obtain the photovoltaic forward adjustable power margin. Multiply the difference between the maximum power point voltage and the current voltage value on the DC side of the photovoltaic system by a preset conversion factor to obtain the photovoltaic negative adjustable power margin. When the photovoltaic-side damping power command is greater than the photovoltaic positive adjustable power margin, the photovoltaic positive adjustable power margin is used; otherwise, the original value is used. Then, the result is compared with the negative value of the photovoltaic negative adjustable power margin for boundary judgment. When the result is less than the negative value of the photovoltaic negative adjustable power margin, the negative value of the photovoltaic negative adjustable power margin is used; otherwise, the original value is used, thus completing the photovoltaic-side command limiting.

10. A method for suppressing low-frequency oscillations in a photovoltaic energy storage system according to claim 8, characterized in that, Limiting of damped power command on the energy storage side includes: The remaining rechargeable capacity is obtained by multiplying the installed energy storage capacity by the current state of charge value, and the remaining dischargeable capacity is obtained by multiplying the difference between the fully charged state value and the current state of charge value by the installed energy storage capacity. When the energy storage side damping power command is greater than the remaining dischargeable capacity, the remaining dischargeable capacity is used; otherwise, the original value is used. Then, the result is compared with the negative value of the remaining rechargeable capacity for boundary judgment. When the result is lower than the negative value of the remaining rechargeable capacity, the negative value of the remaining rechargeable capacity is used; otherwise, the original value is used, thus completing the energy storage side command limiting. Calculate the absolute value of the difference between the energy storage side damping power command after the limit is set and the value of the previous control cycle. When the absolute value of the difference is greater than the preset change limit, the current command is corrected to the algebraic sum of the value of the previous cycle and the preset change limit. The sign of the corrected command is consistent with that of the command before the correction.

Images