Regional damping support control method for wind turbine group in high-proportion new energy power grid

Abstract

The application provides a kind of high proportion new energy power grid network construction wind turbine group area damping support control method, belongs to power system stability control field.On-line identification area leading mode damping ratio and relative oscillation angular velocity, on-line constructs control effectiveness gain of network construction wind turbine group to area leading oscillation channel and checks its control direction;By the difference between the total equivalent damping demand corresponding to the target damping ratio and the intrinsic damping is determined by the gain inversion performance requirement lower bound, according to the bidirectional symmetric active margin and the amplitude of oscillation angular velocity to determine the upper limit of physical ability;When lower bound does not exceed upper limit, close-loop adjustment damping support coefficient in constraint interval makes damping ratio track target, when lower bound exceeds upper limit, limit damping support coefficient to upper limit and output the remaining damping shortage to system level stability control layer, direction check does not pass, stop inversion and exit additional damping channel;Accordingly, additional active power in the direction opposite to the oscillation angular velocity is superimposed to the active reference value of network construction wind turbine group.

Description

Technical Field

[0001] This invention belongs to the field of power system stability control technology, and particularly relates to a regional damping support control method for wind turbine groups in high-proportion renewable energy power grids. Background Technology

[0002] With the large-scale grid connection of new energy power generation such as wind power and photovoltaic power, the proportion of conventional synchronous power sources in the power grid continues to decline, and the equivalent inertia and damping level of the system decrease accordingly. The damping level of low-frequency oscillations between regions is closely related to the operating mode, the start-up combination, and the output level of new energy sources. The weak damping phenomenon is more prominent in power grids with a high proportion of new energy sources, which may threaten the safe and stable operation of the power grid.

[0003] Grid-connected wind turbines, by simulating the rotor motion and voltage regulation characteristics of synchronous generators, can actively establish voltage and frequency, and achieve rapid active and reactive power regulation with the help of converters, providing a means for damping control of regional low-frequency oscillations. Existing technologies for using grid-connected equipment with additional damping control to suppress low-frequency oscillations mainly include: constructing additional damping channels through bandpass filtering and lead / lag phase compensation, and setting the control gain based on residue, dominant pole configuration, or grid strength indicators; superimposing signals such as grid connection point voltage amplitude deviation, phase angle deviation, tie-line interactive power, or power derivative onto the power control loop through feedback gain, and improving dynamic characteristics by adjusting equivalent damping; optimizing the layout and capacity configuration of grid-connected equipment based on indicators such as modal participation factor and stability margin; and using the regional dominant mode damping ratio as the closed-loop regulation target to suppress inter-regional oscillations by changing the output power of power electronic resources.

[0004] The above methods have the following shortcomings in practical applications of high-proportion renewable energy power grids:

[0005] 1. The control effect of grid-type wind turbine clusters on the regional dominant oscillation mode is closely related to the connection location, operation mode and the degree of unit participation. Existing methods mostly adjust the control gain of the additional damping to the regional damping ratio based on experience or approximate it as a constant value. The determination of the required damping support lacks online quantitative basis.

[0006] 2. Additional damping control is based on the premise of forming positive damping for the regional mode. When the operating mode changes or the control parameters are mismatched, causing the control action to reverse, the existing method lacks online judgment and corresponding handling means, which may lead to negative damping and aggravate oscillation.

[0007] 3. The regional damping performance requirements vary with the operating mode, while the damping support capacity of the grid-type wind turbine group is constrained by the bidirectional active power regulation margin and conditions such as converter current and thermal stability. Existing methods are mostly limited to simple amplitude limiting or fixed parameter adjustment on the equipment side, which makes it difficult to take into account both the damping performance requirements and the actual adjustment capacity of the equipment.

[0008] 4. When the available adjustment capacity is insufficient to meet the target damping requirements, existing methods lack quantitative assessment of the remaining damping demand and means of coordinating with system-level stability control.

[0009] 5. The online estimation of the regional dominant mode damping ratio is affected by measurement noise and changes in operating mode. The accuracy and reliability of the estimation results directly affect the control performance. Existing methods are still not perfect in terms of the engineering design of the estimation process and the assurance of its reliability.

[0010] Therefore, in a high-proportion renewable energy power grid, how to ensure that grid-connected wind turbine clusters can provide effective and reliable damping support for the regional power grid under different operating modes is a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0011] To address the shortcomings and deficiencies of existing technologies, this invention provides a regional damping support control method for wind turbine clusters in high-proportion renewable energy power grids. This method uses the regional dominant mode damping ratio as the closed-loop controlled variable and the regional equivalent damping support coefficient of the grid-connected wind turbine group as the control variable to construct the control effectiveness gain of the grid-connected wind turbine group to the regional dominant oscillation channel online. This allows the conversion relationship between the damping support coefficient and the regional dominant mode damping ratio to be updated online with the operating status, providing a quantitative basis for the inversion of damping demand. The control direction of the control effectiveness gain is checked online, and passing the check is used as the entry condition for performing damping inversion. When the check passes, the lower bound of performance demand is obtained from the target damping ratio inversion, and the upper bound of physical capacity is constructed from the group's adjustment capability. The damping support coefficient is adjusted in a closed loop using the two-sided constraints to make the regional dominant mode damping ratio track the target value. When the demand exceeds the physical capacity, the maximum safety support is provided according to the physical upper bound, and the remaining damping deficit that still cannot be met is output to the system-level stability control layer. When the check fails, the inversion is stopped and the regional additional damping channel is withdrawn. The damping support coefficient is ultimately decomposed to each grid-connected wind turbine group according to the power allocation weight, generating additional active power that is opposite to the regional relative oscillation angular velocity, which is superimposed on the original grid-connected active power reference value of the turbine group, thereby realizing the damping support of the grid-connected wind turbine group for the regional power grid.

[0012] The present invention provides a method for regional damping support control of wind turbine clusters in high-proportion renewable energy power grids, comprising:

[0013] Online acquisition of estimated damping ratio of dominant mode in the region and relative oscillation angular velocity of the region;

[0014] The control effectiveness gain of the online construction of the grid-connected wind turbine group to the regional dominant oscillation channel, and its product with the damping support coefficient constitutes the additional equivalent damping provided by the grid-connected wind turbine group;

[0015] The control direction of the control effectiveness gain is verified online, including comparing the control effectiveness gain with the data-side gain.

[0016] The lower bound of the performance requirement is obtained by inverting the control effectiveness gain based on the difference between the total equivalent damping requirement of the region corresponding to the target damping ratio and the intrinsic damping of the region estimated online; the upper bound of the physical capability is determined based on the bidirectional symmetrical active power margin of the grid-connected wind turbine group and the amplitude of the relative oscillation angular velocity of the region.

[0017] When the verification is passed and the lower bound of the performance requirement does not exceed the upper bound of the physical capability, the damping support coefficient is adjusted in a closed loop within the constraint range formed by the two to make the estimated damping ratio track the target damping ratio.

[0018] When the verification is passed and the lower bound of the performance requirement exceeds the upper bound of the physical capacity, the damping support coefficient is limited to the upper bound of the physical capacity, the remaining damping deficit is calculated and output to the system-level stability control layer.

[0019] If the verification fails, stop the inversion based on the control effectiveness gain and exit the regional additional damping channel;

[0020] The additional active power generated based on the damping support coefficient is opposite to the regional relative oscillation angular velocity and is superimposed on the original grid active power reference value of the grid-connected wind turbine group.

[0021] The control effectiveness gain characterizes the equivalent damping strength that a unit damping support coefficient can provide for the dominant oscillation mode in the region. Its product with the damping support coefficient constitutes the additional equivalent damping provided by the wind turbine cluster in the grid, and has the same dimensions as the intrinsic damping of the region. It can be directly compared and superimposed with the intrinsic damping of the region, providing a unified benchmark for the conversion between the total equivalent damping demand of the region and the required damping support coefficient. The data-side gain is obtained online by identifying the incremental data of the damping ratio estimate of the dominant mode in the region and the damping support coefficient, reflecting the direction of control action embodied in the actual input-output response. Comparing the control effectiveness gain calculated based on the model with the data-side gain can simultaneously utilize the model information and actual response information at the current operating point, reducing single-source misjudgment caused by model mismatch or measurement anomalies. If the verification fails, the inversion is stopped and the regional additional damping channel is withdrawn. This can prevent the inversion from continuing and forming negative damping when the direction of control action is unreliable. The regional additional damping channel refers to the additional damping channel superimposed on the active power control channel of the grid-connected wind turbine group, which is opposite to the regional relative oscillation angular velocity. After the channel is withdrawn, the basic control functions of the grid-connected wind turbine group in establishing voltage and frequency are not affected.

[0022] Optionally, the regional relative oscillation angular velocity is determined by the difference between the inertia center angular velocities of two main coherent regions in the regional power grid. These two main coherent regions are defined based on the phase characteristics of the dominant regional oscillation mode and the frequency response of the inertia center. The observation vector characterizing the regional relative oscillation angular velocity is constructed according to the inertia center weights of the two main coherent regions. The bidirectional symmetrical active power margin of the grid-connected wind turbine group is taken as the minimum value among the group's upward active power adjustment margin, downward active power adjustment margin, and the corresponding margins of converter current, power, and thermal stability limits. This value serves as the symmetrical margin of the group's usable power for alternating positive and negative damping. The estimated damping ratio of the dominant regional mode can be obtained by constructing a multi-channel measurement vector from the regional relative oscillation angular velocity, the regional equivalent relative work angle, and the active power of the tie lines between the main coherent regions. The regional dominant oscillation mode eigenvalues ​​are identified using the multi-channel overall least squares matrix pencil method, and the estimated damping ratio is obtained from these eigenvalues. Furthermore, a damping ratio estimation credibility indicator can be generated based on the signal-to-noise ratio, modal continuity, and multi-channel consistency. When the damping ratio estimation credibility indicator indicates that the estimation is unreliable, the accumulation of the integral element is paused, and the current damping support coefficient is maintained or reduced at a slow release rate.

[0023] Further, the lower bound of the performance requirement is determined as follows: the total equivalent damping requirement of the region is calculated based on the target damping ratio, the estimated value of the regional equivalent inertia, and the estimated value of the regional equivalent synchronous stiffness; the difference between the total equivalent damping requirement of the region and the intrinsic damping of the region is divided by the control effectiveness gain, and the larger value between the result and zero is taken as the lower bound of the performance requirement; the intrinsic damping of the region is obtained online by inverse solving based on the damping ratio estimate, the estimated value of the regional equivalent inertia, the estimated value of the regional equivalent synchronous stiffness, the control effectiveness gain, and the current damping support coefficient, according to the mapping relationship between the regional dominant mode damping ratio and the total equivalent damping of the region. The mapping relationship between the regional dominant mode damping ratio and the regional total equivalent damping can be determined by the second-order equivalent model of the regional dominant oscillation. The estimated damping ratio reflects the closed-loop damping ratio after control is implemented, which includes the additional equivalent damping provided by the control implementation. Subtracting this part yields the regional intrinsic damping before control, allowing the lower bound of performance requirements to be updated in a timely manner with changes in operating mode. This inverse solution can be performed during periods when the damping ratio estimate is reliable and the change in the damping support coefficient within the identification window does not exceed a set threshold. If necessary, the inverse solution can be performed separately and cross-checked during two quasi-steady-state periods with different damping support coefficient levels. When the obtained result is non-positive, it is taken as zero, indicating that the regional intrinsic damping has met the target requirements, and the fleet does not need to implement additional damping support.

[0024] Furthermore, the control effectiveness gain of constructing the grid-connected wind turbine group to the regional dominant oscillation channel online includes: calculating the sensitivity of the regional dominant oscillation mode eigenvalues ​​to the damping support coefficient based on the linearized state matrix of the current operating point of the regional power grid, the weighted equivalent input matrix of the grid-connected wind turbine group, the right and left eigenvectors of the regional dominant oscillation mode, and the observation vector characterizing the regional relative oscillation angular velocity; and multiplying the sensitivity by twice the square root of the product of the regional equivalent inertia and the regional equivalent synchronous stiffness after the damping ratio sensitivity transformation to obtain the control effectiveness gain. The weighted equivalent input matrix of the grid-connected wind turbine group can be obtained by weighted summation of the additional active power input matrices of each grid-connected wind turbine group according to the power distribution weights; the sensitivity of the eigenvalue to the damping support coefficient can be determined by the first-order eigenvalue perturbation relationship, reflecting the change of the dominant mode eigenvalue caused by the change of the damping support coefficient, and transformed into the damping ratio sensitivity through the conversion relationship between the damping ratio and the eigenvalue; multiplied by twice the square root of the product of the regional equivalent inertia and the regional equivalent synchronous stiffness, making the gain dimensionless with the regional equivalent damping, and its product with the damping support coefficient can be directly compared and superimposed with the regional intrinsic damping; the gain is updated online with changes in the operating point, connection location, and participating turbine group.

[0025] Furthermore, the single-unit control effectiveness gain of each candidate grid-connected wind turbine group is calculated separately. Grid-connected wind turbine groups with single-unit control effectiveness gains greater than a set threshold constitute an effective turbine group set. Based on the bidirectional symmetrical active power margin and single-unit control effectiveness gain of each grid-connected wind turbine group within the effective turbine group set, the comprehensive support capability index of each grid-connected wind turbine group is calculated. The power allocation weights of each grid-connected wind turbine group are normalized and then weighted and aggregated according to the power allocation weights to obtain the control effectiveness gain. The single-unit control effectiveness gain can be calculated using the additional active power input matrix of each turbine group, following the aforementioned construction method. An entry threshold is set to filter out turbine groups with weak effects on the regional dominant oscillation mode. Its value can be determined based on the minimum control action intensity required by the turbine group for the current dominant mode, combined with simulation calculations and field tests. The comprehensive support capability index considers both the control action intensity and available adjustment capability of the turbine group, ensuring that turbine groups with significant control action and large adjustment margins occupy a higher proportion in power allocation. Optionally, when the effective fleet is empty, the input of additional damping in the region is suspended, the current damping support coefficient is maintained or it is released to zero according to the safe release rate, and then re-screened after the operating mode changes.

[0026] Furthermore, the online verification includes: calculating the model-side control effectiveness gain from the linearized model of the current operating point of the regional power grid; establishing a local autoregressive model based on the data sequence consisting of the estimated damping ratio and the increment of the damping support coefficient, and obtaining the data-side gain through recursive least squares online identification; within a continuously set number of update cycles, when both the model-side control effectiveness gain and the data-side gain are positive, and the relative deviation between the data-side gain and the local control gain determined by the model-side control effectiveness gain, the regional equivalent inertia, and the regional equivalent synchronous stiffness does not exceed the set tolerance, the control direction is determined to be reliable; otherwise, the control direction is determined to be unreliable. The positive values ​​of both the model-side control effectiveness gain and the data-side gain indicate that both model calculations and actual responses suggest that increasing damping support helps improve the regional damping level. The local control gain represents the conversion relationship between the damping support coefficient and the regional dominant mode damping ratio, and is the quotient obtained by dividing the model-side control effectiveness gain by twice the square root of the product of the regional equivalent inertia and the regional equivalent synchronous stiffness. The relative deviation between the data-side gain and the local control gain is used to verify the consistency of their amplitudes and avoid misjudgments caused by magnitude mismatch. The set tolerance can be, for example, 30% to 50%, and the continuous set number can be, for example, three to five update cycles. The specific values ​​can be determined by combining simulation calculations and field tests. Correct control direction is a prerequisite for stable operation of closed-loop regulation: under small-signal conditions that ignore the effects of saturation and amplitude limiting, the dynamics of the damping ratio tracking error can be approximated as a second-order homogeneous equation. When the actual direction of the control action on the regional oscillation channel is consistent with the direction based on the controller, the error dynamically converges; when the direction is reversed, the error dynamically exhibits a positive feedback divergence trend. Optionally, before the initial implementation of regional additional damping control or before its reinstatement after withdrawal, a limited-amplitude trial damping support command can be applied to the grid-connected wind turbine group, provided that the bidirectional symmetrical active power margin and rate of change of each grid-connected wind turbine group are met, in order to collect the data sequence required for recursive least squares. During the trial period, normal closed-loop control is not entered.

[0027] Furthermore, the upper bound of the physical capability is determined as follows: the square of the relative oscillation angular velocity of the region is subjected to a first-order recursive filter and the square root is taken to obtain the slow-varying root mean square amplitude. The larger value between the slow-varying root mean square amplitude and the minimum amplitude threshold is taken as the effective oscillation amplitude. Based on the bidirectional symmetrical active power margin, power distribution weight and effective oscillation amplitude of each grid-connected wind turbine group within the effective turbine group set, the upper limit of the damping support coefficient of each grid-connected wind turbine group is calculated. The minimum value is compared with the hardware upper limit of the damping coefficient, and the smaller value is taken as the upper bound of the physical capability. The additional damping power provided by the generator group varies with the relative oscillation angular velocity of the region, alternating between positive and negative. Its available adjustment capability is constrained by the most stringent of three conditions: upward adjustment, downward adjustment, and converter capability. The square root of the square of the relative oscillation angular velocity of the region is obtained by performing a first-order recursive filter, which is a slowly varying root mean square amplitude. This avoids upper limit jitter caused by converting instantaneous changes. Setting a minimum amplitude threshold can prevent abnormal amplification of the upper limit calculated when the regional oscillation amplitude is very small. Each generator group is calculated according to the power allocation weight and effective oscillation amplitude to obtain its own upper limit of damping support coefficient. The minimum value represents the most limited support capability on the generator group side. This value is then compared with the upper limit of the damping coefficient hardware and the smaller value is taken. This indicates that the upper limit of physical capability is constrained by both the generator group adjustment capability and the hardware capability of the control device.

[0028] Further, the closed-loop adjustment damping support coefficient includes: using the lower bound of the performance requirement as the damping requirement feedforward; calculating the difference between the target damping ratio and the estimated damping ratio as the tracking error; performing a proportional-integral operation on the tracking error, dividing the result of the proportional-integral operation by the local control gain, and then superimposing it on the damping requirement feedforward to obtain the nominal damping control command; the local control gain is determined by the control effectiveness gain, the regional equivalent inertia, and the regional equivalent synchronous stiffness; projecting the nominal damping control command onto the interval formed by the lower bound of the performance requirement and the upper bound of the physical capability, applying an asymmetric rate of change constraint to the projection result, wherein the increase rate of the damping support coefficient is greater than the release rate, to obtain the actually executed damping support coefficient; and feeding back the difference between the actually executed damping support coefficient and the nominal damping control command, after conversion by the local control gain, to the integral stage of the proportional-integral operation. Specifically, the lower bound of performance requirements is used as the feedforward quantity, ensuring that the damping support coefficient starts from the minimum support required to meet the target requirements, and the tracking error only needs to be adjusted within a small range; the result of the proportional-integral operation is divided by the local control gain to match the conversion relationship between the feedback correction quantity and the damping support coefficient to the damping ratio of the dominant mode in the region; the projection operation ensures that the nominal damping control command does not exceed the range formed by the two-sided constraints; the increase rate is greater than the release rate, so that the damping support is put into operation as soon as possible when the damping is insufficient and released slowly after stabilization, suppressing frequent adjustments near the constraint boundary; the difference between the actual executed damping support coefficient and the nominal damping control command is converted by the local control gain and fed back to the integral link to keep the integral state consistent with the actual execution state, suppressing integral saturation during the constraint period.

[0029] Furthermore, when the lower bound of performance requirements exceeds the upper bound of physical capability, the actual damping support coefficient is taken as the upper bound of physical capability, and the integral state is calculated back based on the actual damping support coefficient, damping requirement feedforward, and tracking error; the remaining damping deficit is the larger of the difference between the total equivalent damping requirement of the region, the intrinsic damping of the region, the product of the control effectiveness gain and the upper bound of physical capability, and zero; the maximum damping ratio that can be achieved after the upper bound of physical capability is implemented is calculated, and the larger of the difference between the target damping ratio and the maximum damping ratio and zero is taken as the remaining damping ratio gap; the infeasibility state flag, the remaining damping deficit, and the remaining damping ratio gap constitute the system-level stability support request information, which is output to the system-level stability control layer. Among them, if the lower bound of performance requirement exceeds the upper bound of physical capability, it indicates that the target damping exceeds the adjustment capability of the fleet under the current operating conditions. In this case, execution is based on the upper bound of physical capability to ensure that the damping support does not exceed the physical constraints of the equipment. The inverse calculation of the integral state can avoid the continued accumulation of unachievable damping requirements in the integral stage. The remaining damping deficit is quantified from the perspective of regional equivalent damping and the remaining damping ratio gap is quantified from the perspective of damping ratio. The damping requirements that are still not met after the fleet has put all available capabilities into operation are quantified and output to the system-level stability control layer along with the infeasibility status flag. This provides quantitative input for the system-level stability control layer to coordinate other stability control resources.

[0030] Furthermore, the additional active power is decomposed according to the power allocation weight of each grid-connected wind turbine group. Each grid-connected wind turbine group generates an additional damping active power command that is opposite to the regional relative oscillation angular velocity based on the allocation result. This command is then superimposed on the original grid-connected active power reference value of the wind turbine group after local power limiting. The local power limiting is based on the bidirectional symmetrical active power margin of each grid-connected wind turbine group, ensuring that the additional damping active power of each group does not exceed its actual adjustment capacity. The superposition of the additional damping active power on the original grid-connected active power reference value ensures that regional damping support is implemented through the power control channel of the turbine group without altering the original grid-connected control structure. The additional active power, which is opposite to the regional relative oscillation angular velocity, consumes oscillation energy during regional oscillation, thus enabling regional damping support to act on the regional dominant oscillation channel.

[0031] Furthermore, the present invention also provides a regional damping support control device for a grid-connected wind turbine cluster in a high-proportion renewable energy power grid, comprising: an estimation module for online acquisition of the estimated value of the regional dominant mode damping ratio and the regional relative oscillation angular velocity; a gain construction module for online construction of the control effectiveness gain from the grid-connected wind turbine cluster to the regional dominant oscillation channel, and online verification of the control direction of the control effectiveness gain; a constraint generation module for determining the lower bound of the performance requirement based on the difference between the regional total equivalent damping requirement corresponding to the target damping ratio and the online estimated regional intrinsic damping, through the inversion of the control effectiveness gain, and determining the upper bound of the physical capability based on the bidirectional symmetrical active power margin of the grid-connected wind turbine cluster and the amplitude of the regional relative oscillation angular velocity; and a monitoring module. The control module is used to stop the inversion based on the control effectiveness gain and exit the regional additional damping channel when the online verification fails. When the online verification passes and the lower bound of the performance requirement does not exceed the upper bound of the physical capacity, it adjusts the damping support coefficient in a closed loop within the constraint interval formed by the two to make the estimated damping ratio track the target damping ratio. When the online verification passes and the lower bound of the performance requirement exceeds the upper bound of the physical capacity, it limits the damping support coefficient to the upper bound of the physical capacity, calculates the remaining damping deficit, and outputs it to the system-level stability control layer. The execution module is used to generate additional active power in the opposite direction to the regional relative oscillation angular velocity based on the damping support coefficient and superimpose it on the original grid active power reference value of the wind turbine group. The functions of each module in the above device correspond to the corresponding links in the aforementioned method, and the specific implementation methods in the aforementioned method are also applicable to this device.

[0032] Compared to existing technologies, this invention and its preferred embodiment, through online construction of equivalent control effectiveness gain, allow the intensity of the control effect of the grid-connected wind turbine group on the regional dominant oscillation channel to be calculated online according to the operating state and compared with the regional intrinsic damping in the same dimension, thus providing a quantitative basis for determining the damping support strength that matches the current operating conditions. The regional intrinsic damping can be obtained by online inverse solution of closed-loop damping ratio estimation, enabling the lower bound of performance requirements to be automatically updated with changes in intrinsic damping, without the need for offline readjustment according to changes in operating mode. Through control direction verification on the model side and data side, the reliability of the control direction is used as the admission condition for performing damping inversion, and the direction is not unreliable. The system stops inversion and withdraws the regional additional damping channel, reducing the risk of negative damping caused by directional misjudgment without affecting the original network control of the fleet. The dual constraints formed by the lower bound of performance requirements and the upper bound of physical capabilities enable online matching between the target damping requirements and the equipment adjustment capabilities. When the capabilities are insufficient, the system executes according to the upper bound of physical capabilities and suppresses integral accumulation, avoiding the continuous accumulation of unrealizable damping requirements. The quantitative output of the remaining damping deficit provides a means for the system-level stability control layer to coordinate other stability control resources. The constraint determination of the regional layer and the power execution of the fleet layer are separated in time scale, which facilitates the connection and implementation with the wide-area measurement system and the station control system. Attached Figure Description

[0033] Figure 1 This is a schematic diagram of the regional dominant mode dual-sided constraint damping closed-loop control principle provided in an embodiment of the present invention;

[0034] Figure 2 A flowchart illustrating the construction and direction determination of the regional control effectiveness gain of a grid-connected wind turbine cluster, provided in an embodiment of the present invention.

[0035] Figure 3 A schematic diagram of the time-varying feasible region and constraint conflict with bilateral constraints provided in an embodiment of the present invention;

[0036] Figure 4 This is a diagram showing the state transition of regional damping support monitoring and control provided in an embodiment of the present invention.

[0037] Figure 5 This is a closed-loop response diagram of the region-dominant mode damping ratio provided in an embodiment of the present invention;

[0038] Figure 6 This is a schematic diagram of the double-sided constraint boundary and the actual damping control quantity provided in an embodiment of the present invention;

[0039] Figure 7 The diagram shows the regional relative oscillation velocity response under different control methods provided in the embodiments of the present invention.

[0040] Figure 8 A schematic diagram of infeasible deficit output and control direction supervision provided for embodiments of the present invention;

[0041] Figure 9 This is a region-dominant oscillation energy decay envelope diagram provided for an embodiment of the present invention. Detailed Implementation

[0042] To make the features and advantages of the present invention more apparent and understandable, specific embodiments are described below in conjunction with the accompanying drawings:

[0043] I. Overall Plan

[0044] This embodiment provides a regional damping support control method for grid-connected wind turbine clusters in high-proportion renewable energy power grids. This method uses the regional dominant mode damping ratio as the closed-loop controlled variable and the regional equivalent damping support coefficient of the grid-connected wind turbine cluster as the control variable. Through the coordination of control effectiveness calculation, control direction determination, feasibility assessment of bilateral constraints, constraint closed-loop implementation, and handling of infeasible states, closed-loop control of regional damping support is achieved. Its control objective is: under the condition that the control direction of the grid-connected wind turbine cluster is correct and the physical adjustment capability allows, to make the regional dominant mode damping ratio track a given target; when the target exceeds the achievable range of the wind turbine cluster, constraint conflicts are identified online, and the remaining damping deficit that the wind turbine cluster cannot bear is quantified as a coordinating input for system-level stability control.

[0045] The overall principle is as follows Figure 1 As shown, the wide-area measurement stage provides regional dynamic measurements for the dominant mode observation stage; the dominant mode observation stage estimates the damping ratio of the regional dominant mode online and provides an estimate confidence indicator; the control effectiveness stage uses computer clusters to control effectiveness gain and performs direction verification; the two-sided constraint generation stage provides the lower bound of performance requirements and the upper bound of physical capabilities; the feasibility monitoring stage provides the two-sided constraint feasibility margin and operating status; the constraint closed-loop control stage generates damping support coefficient instructions according to feedforward, feedback, projection, and anti-integral saturation; the cluster power allocation stage decomposes the instructions to each cluster, executes them locally, and superimposes them onto the original network active power reference; the response of the regional power grid dominant oscillation channel is fed back to the measurement and mode observation stages, forming a closed loop. When constraint conflicts or control anomalies occur, the infeasibility and anomaly handling stage outputs the remaining damping deficit, triggers a safe exit, and provides coordination information to the system-level stability control layer.

[0046] For ease of description, scalar quantities are represented by italic letters, and vectors and matrices are represented by bold; variables with superscript caps indicate online estimates; superscripts ref, lim, and hw represent reference values, operational limits, and hardware limits, respectively; subscript R indicates quantities related to the regional dominant oscillation, subscript W indicates quantities related to the grid-connected wind turbine cluster, and subscript... Indicates the first Individual machine clusters; It represents the increment of the small signal or the change in adjacent update cycles. It is a non-negative truncation operator; To be Projected onto closed interval The projection operator; To be Limited to The saturation operator within; the superscripts T and H represent the transpose and conjugate transpose, respectively.

[0047] II. Regional Dominant Oscillation Control Model

[0048] Consider the small-signal model of the regional power grid under the current operating mode, as shown in Equation (1).

[0049] (1)

[0050] In the formula, This is the regional power grid state increment vector; This is the linearized state matrix at the current running point; The number of candidate grid-connected wind turbine clusters; For the first Additional active power input matrix for each cluster; For the first Additional active power of each generator group; The input matrix is ​​the external perturbation matrix; The vector represents external disturbances such as load switching and changes in renewable energy output; the upper point indicates the derivative with respect to time.

[0051] Suppose that the dominant oscillation mode of the current restricted area's stable operation capability is as shown in equations (2) to (4).

[0052] (2)

[0053] (3)

[0054] (4)

[0055] In the formula, These are the eigenvalues ​​of the dominant regional oscillation mode; The real part of the eigenvalue; This is the damped oscillation angular frequency; The imaginary unit; This is the dominant oscillation frequency in the region; , These are the equivalent work angles of the inertia centers of the two main coherent regions A and B, respectively, divided according to the dominant mode phase and frequency response. , These are the angular velocities of the inertia centers of the two coherent regions, respectively. This is the equivalent relative work angle for the region; The relative angular velocity of the region.

[0056] The relative power angle and relative oscillation angular velocity of the region satisfy equation (5). Near the dominant mode, the higher-order model of the regional power grid is projected into a second-order equivalent oscillation model, as shown in equation (6); the wind turbine group of the grid adopts an additional active power damping input that is opposite to the regional relative oscillation angular velocity, as shown in equation (7).

[0057] (5)

[0058] (6)

[0059] (7)

[0060] In the formula, The equivalent inertia of the relative motion between the two coherent regions; The intrinsic equivalent damping is when no damping control is applied in the wind power grid area; For the equivalent synchronous stiffness of the region; The equivalent control effectiveness gain from the grid-connected wind turbine cluster to the regional dominant oscillation channel; For area damping control input; This is the equivalent external disturbance projected onto the oscillation channel; The equivalent damping support coefficient for the grid-connected wind turbine cluster area is specified. .

[0061] Substituting equation (7) into equation (6), we can obtain the second-order equivalent model of the region after applying control, as shown in equation (8); the ratio of the total equivalent damping of the region to the damping ratio of the dominant mode of the region are shown in equation (9) and equation (10), respectively.

[0062] (8)

[0063] (9)

[0064] (10)

[0065] In the formula, The total equivalent damping of the region after grid-connected wind power participation, where Additional equivalent damping provided to the grid-connected wind turbine clusters through regional additional damping channels; The damping ratio of the dominant mode in the region; The undamped natural angular velocity of the equivalent regional model is given by equation (10). Equation (10) establishes the basic control mapping from the damping support coefficient to the damping ratio of the dominant regional mode, which serves as the basis for subsequent performance requirement inversion and control design.

[0066] III. Construction of Control Effectiveness Gain and Online Verification of Control Direction

[0067] To avoid simply equating the control effect of the grid-connected wind turbine cluster on the regional dominant oscillation channel to a constant, this method constructs the equivalent control effectiveness gain as an online computable quantity related to the dominant mode, grid connection location, and turbine cluster participation weight. Its construction and direction determination process are as follows: Figure 2 As shown. The relative oscillation angular velocity of the region can be obtained by observing the state increment vector, as shown in equation (11).

[0068] (11)

[0069] In the formula, This represents the observation vector of the relative angular velocity of the region's oscillations. As a preferred implementation... It can be constructed based on the relationship between the inertia center weights and angular velocity observations of two coherent regions, and updated according to the coherent region division results.

[0070] The additional damping active power of each group of machines is generated according to the regional damping support coefficient and the allocation weight, as shown in Equations (12) and (13); the weighted equivalent input matrix of the group of machines is shown in Equation (14).

[0071] (12)

[0072] (13)

[0073] (14)

[0074] In the formula, For the first Regional damping power allocation weights for individual machine groups; The weighted equivalent input matrix for the wind turbine cluster.

[0075] The closed-loop state matrix of the regional power grid after control is applied is shown in Equation (15). The first-order sensitivity of the regional dominant mode eigenvalue to the damping support coefficient is shown in Equation (16), and can be decomposed into real part sensitivity and angular frequency sensitivity, as shown in Equation (17).

[0076] (15)

[0077] (16)

[0078] (17)

[0079] In the formula, This is the closed-loop state matrix after applying damping control to the grid-connected wind power area; The complex sensitivity of the regional dominant mode eigenvalues ​​to the damping support coefficient; , These are the right and left eigenvectors of the dominant mode, respectively. , These are the real part of the eigenvalue and the sensitivity of the damped oscillation angular frequency to the damping support coefficient, respectively.

[0080] The relationship between the damping ratio of the dominant mode and the eigenvalue is shown in Equation (18). The sensitivity of the damping ratio to the damping support coefficient can be determined by Equation (19) according to the chain relationship. Thus, the equivalent control effectiveness gain of the grid-connected wind turbine group to the dominant oscillation channel is defined as shown in Equation (20).

[0081] (18)

[0082] (19)

[0083] (20)

[0084] In the formula, This is the local sensitivity of the region-dominant mode damping ratio to the damping support coefficient. Equation (20) is obtained by multiplying by... Make the control effectiveness gain dimensionless with the region equivalent damping, thereby enabling Can be directly connected to intrinsic damping Adding them together provides a unified dimensional basis for performance requirement inversion and physical capability comparison.

[0085] The control effectiveness gain of each cluster is constructed in the same way, as shown in Equations (21) and (22); based on this, the cluster set participating in the current dominant mode damping support is selected, as shown in Equation (23).

[0086] (twenty one)

[0087] (twenty two)

[0088] (twenty three)

[0089] In the formula, For the first Sensitivity of individual machine group damping control to the regional dominant mode characteristic value; For the first Control effectiveness gain for individual clusters; For the reason Damping ratio sensitivity obtained by the same transformation according to equation (19); To form a fleet of aircraft participating in the current regional dominant mode of damping support; This is the minimum control effectiveness admission threshold.

[0090] The comprehensive support capability index and allocation weight of the fleet are shown in Equations (24) and (25), and the overall control effectiveness gain of the fleet is shown in Equation (26).

[0091] (twenty four)

[0092] (25)

[0093] (26)

[0094] In the formula, For the first The comprehensive regional damping support capacity index of an individual machine group; The bidirectional symmetrical active power margin of this generator group is described in detail below. To participate in the overall control effectiveness gain of the fleet, the power allocation weight of fleets that are not included in the effective fleet set is set to zero.

[0095] As a preferred implementation, if the set If the value is empty, the additional damping channel in the region will not be activated in this control cycle. The current damping support coefficient will be maintained or released to zero at a safe rate, and a flag indicating that the set is empty will be output. The access screening will be re-executed after the operating mode changes.

[0096] To ensure the correctness of the direction of the damping inversion, this method performs online verification of the control direction. A local autoregressive model is established using the incremental sequence of the damping ratio estimate and the damping support coefficient, as shown in Equation (27); the confidence index of the control direction is determined according to Equation (28).

[0097] (27)

[0098] (28)

[0099] In the formula, For discrete sampling sequence numbers; is the local autoregressive coefficient of the regional damping ratio; The data-side gain is calculated based on the change in damping support coefficient to the change in damping ratio. The data-side gain is obtained by online estimation using recursive least squares. The equivalent discrete delay is measured in update cycles; Unmodeled dynamics and measurement noise; , These represent the increments of the regional dominant mode damping ratio estimate and the damping support coefficient, respectively, within adjacent update cycles; As a reliable indicator for control direction; The control effectiveness gain calculated on the model side; This is the minimum validity threshold for the model side; The threshold for determining the direction of the data; To determine the number of consecutive confirmation cycles. When the condition in equation (28) is not met, After verification, the controller adopts the control effectiveness gain. Take the calculated values ​​from the model side This design transforms the theoretical assumption of correct control direction into a practical closed-loop admission condition, provided that... Then stop gain inversion based on control effectiveness gain to avoid negative damping or positive feedback caused by sign judgment error.

[0100] As a preferred implementation, the data-side gain and model-side results can also be checked for amplitude consistency: the local gain from the damping ratio to the damping support coefficient is calculated from the model-side gain. ,when The direction is considered reliable only when the relative deviation from the converted value does not exceed a given tolerance, for example, 30% to 50%; the number of consecutive confirmation periods. For example, take three to five times; data side direction discrimination threshold. It can be set at 0.3 to 0.5 times the converted value and adjusted in conjunction with on-site tests.

[0101] As a preferred implementation method, during the re-entry phase after the initial deployment or safe withdrawal, a trial damping support command with limited amplitude and rate can be applied first, such as a small step or a low-frequency small-amplitude sine signal. Under the premise of meeting the power limit and rate of change limit of the generator group, the damping ratio increment and damping support coefficient increment sequence are collected for use in recursive least squares estimation. During the trial period, the normal closed loop is not entered. After the trial, the decision on whether to deploy is made based on the direction verification results.

[0102] As a preferred implementation, the model-side gain is calculated on the small-signal model at the current operating point. The state matrix and input matrix can be obtained from the state estimation results of the energy management system and the station parameter library, or obtained through online identification. The dominant mode eigenvalues ​​and left and right eigenvectors are obtained from state matrix feature analysis and cross-checked with the frequencies and damping ratios obtained from online mode identification. When changes in operating mode cause significant changes in modal parameters or participating weights, the model-side gain is recalculated, and its calculation cycle can be consistent with the damping ratio update cycle. Considering that the damping support coefficient level will change the regional modal eigenvalues ​​and modal eigenvectors, sensitivity is prioritized in the closed-loop state matrix after considering the current damping support level. The values ​​are obtained at the current location and can be updated iteratively with the damping support coefficient. Specifically, the sensitivity and single-unit group gain are first calculated based on the current values. After updating the damping support coefficient, the modal eigenvectors and sensitivity are recalculated. After one or two iterations, this is used as the model-side gain for this update cycle to reduce extrapolation errors caused by the sensitivity point deviating from the current operating state. In cases where the damping support coefficient is small, it can also be directly obtained from the state matrix at the current operating point. The sensitivity is calculated and its applicable operating range is limited.

[0103] IV. Online Estimation of Damping Ratio of the Dominant Mode in the Region

[0104] The regional dominant mode damping ratio is the feedback quantity of the closed-loop control, and its online acquisition adopts the multi-channel overall least squares matrix pencil method. The multi-channel measurement vector is shown in Equation (29), the identified modal eigenvalues ​​and damping ratio estimates are shown in Equations (30) and (31), and the estimation confidence flag is shown in Equation (32).

[0105] (29)

[0106] (30)

[0107] (31)

[0108] (32)

[0109] In the formula, For the first Multi-channel region dynamic measurement vector at each sampling time; The active power of the connecting lines between the main coherent areas; These are the regional dominant mode feature values ​​obtained through online identification; , These represent the real part of the mode and the damped oscillation angular frequency, respectively, estimated online. This is an online estimate of the damping ratio of the dominant mode in the region; The damping ratio estimation reliability flag is set to 1 when the signal-to-noise ratio, modal continuity, and multi-channel consistency all meet the threshold requirements, and otherwise to 0.

[0110] Recommended online modal estimation parameters are shown in Table 1. Under low signal-to-noise ratio conditions, the reliability of the damping ratio estimation decreases. In this case, the integral should be frozen and the damping support coefficient maintained or gradually released to prevent the unreliable feedback-driven control quantity from continuing to increase.

[0111] Table 1 Recommended Online Modality Estimation Parameters

[0112]

[0113] Online estimation obtained The closed-loop damping ratio is given after applying regional damping control. To serve the performance requirement inversion, the regional intrinsic damping can be obtained online according to equation (33).

[0114] (33)

[0115] In the formula, This is an online estimate of the region's intrinsic damping; , These are the online estimates of the regional equivalent inertia and synchronous stiffness, respectively, which can be updated near the current operating point through model projection, parameter inversion, or recursive estimation. The gain for online control effectiveness adopted by the controller. As a preferred implementation, equation (33) estimates the confidence flag. Furthermore, it is used to identify periods within the window where the damping support coefficient remains approximately constant; when there is a significant change in the damping support coefficient within the window, it can be combined with the incremental model of equation (27) to isolate the contribution of the control action to the change in the damping ratio before calculating the intrinsic damping; if necessary, the intrinsic damping can be extracted and cross-checked in two quasi-steady-state periods with different damping support coefficient levels. The intrinsic damping estimate is refreshed according to the damping ratio update cycle and maintained between adjacent updates for use in the calculation of the lower bound of performance requirements and feedforward quantities.

[0116] V. Construction and Feasibility Assessment of Two-Sided Constraints

[0117] The regional damping closed-loop reference value is determined by the minimum allowable damping ratio plus a safety margin, as shown in Equation (34); the corresponding regional total equivalent damping requirement is shown in Equation (35).

[0118] (34)

[0119] (35)

[0120] In the formula, This is the reference value for the closed-loop damping of the region; The minimum allowable damping ratio for the dominant mode in the region; To provide a safety margin to cover estimation errors and operational fluctuations; The total equivalent damping of the region required to achieve the target damping ratio.

[0121] The minimum wind resistance damping support coefficient required to meet the target damping requirements can be obtained by inversion from equation (10), which is the lower bound of the performance requirement, as shown in equations (36) and (37).

[0122] (36)

[0123] (37)

[0124] In the formula, The minimum structural wind resistance damping support coefficient required to meet the target damping requirements. As can be seen from the non-negative truncation, the lower bound is zero when the intrinsic damping already meets the reference requirements, at which point no additional damping support is needed; The inversion is contingent upon the direction verification being passed. The positive value is guaranteed by the admission conditions. Equation (37) is equivalent to Equation (36), and directly gives the inversion relationship between the target damping ratio and the lower limit of the wind power control input.

[0125] The upper limit of physical capability is determined by the two-way symmetrical active power margin of the cluster, the slow amplitude of the regional oscillation, and the upper limit of the hardware. The two-way symmetrical active power margin of the cluster is shown in equation (38).

[0126] (38)

[0127] In the formula, , The first Individual generator groups can adjust the available margin of active power upwards and downwards; This refers to the available active power margin corresponding to the converter current, power, and thermal stability limitations. The minimum value among the three is used as the bidirectional symmetrical active power margin for the generator group that can be used for alternating positive and negative damping power.

[0128] The slowly varying root mean square amplitude of the relative oscillation angular velocity of the region is updated recursively according to equation (39), and the effective oscillation amplitude used for physical upper bound calculation is taken as the larger of the filter value and the lower limit threshold according to equation (40).

[0129] (39)

[0130] (40)

[0131] In the formula, This represents the slowly varying root-mean-square amplitude of the relative oscillation angular velocity of the region. For amplitude filtering update coefficients; The minimum amplitude threshold is set to prevent abnormal amplification of the upper bound under low amplitude conditions; The effective oscillation amplitude is used for calculating the physical upper bound. The initial value of the recursion in equation (39) can be the root mean square value of the relative oscillation angular velocity within the first sliding window, or it can be zeroed and then converged quickly by the recursive relationship. This process avoids directly using fast variables with zero-crossing periods as divisors for upper bound calculations, thus suppressing the risk of upper bound jitter and division by zero.

[0132] The additional damping active power amplitude of each group of generators under the regional oscillation amplitude should not exceed its bidirectional symmetrical active power margin, as shown in Equation (41); based on this, the maximum achievable grid wind resistance damping support coefficient under the current physical conditions is obtained, i.e., the upper limit of physical capacity, as shown in Equation (42).

[0133] (41)

[0134] (42)

[0135] In the formula, The maximum achievable wind resistance damping support coefficient for the network under current physical conditions; This represents the upper limit of the damping coefficient allowed by the controller and device. Under the above weighting rules, the constraint bottleneck is determined by the group of machines with the largest control effectiveness gain. The resulting physical upper bound and the residual damping deficit calculated based on this upper bound are both conservative safety-side estimates.

[0136] As a preferred implementation, the slow amplitude value in equation (39) is statistically calculated using the root mean square (RMS) caliber, while the peak value of the instantaneous additional damping active power of each generator group is approximately 1.4 times the RMS amplitude under near-sinusoidal oscillation. To balance the coordination between the support capacity exploitation and the execution layer limit, a crest coefficient can be introduced into equations (41) and (42), i.e., a power margin is reserved based on the peak value after multiplying the RMS amplitude by the crest coefficient; if the crest coefficient is not introduced, the execution layer limit will naturally be clipped, and the actual effective additional damping will be slightly lower than the model prediction value, which can be considered in parameter tuning and simulation verification.

[0137] The feasibility margin of the two-sided constraints and its state determination are shown in equations (43) to (46).

[0138] (43)

[0139] (44)

[0140] (45)

[0141] (46)

[0142] In the formula, This represents the feasibility margin under both-sided constraints. The safety margin threshold reserved between normal feasibility and critical feasibility can be determined by the damping ratio estimation error and the typical amplitude of operational fluctuations. The flag for infeasibility is 1 when equation (46) is true, and 0 otherwise. Equation (44) corresponds to the normal feasible state, where the target damping can be achieved under sufficient margin conditions; Equation (45) corresponds to the critical feasible state, where the target damping can still be achieved but the damping resources are close to the physical limit, at which point the controller suppresses the integral element and outputs a warning; Equation (46) corresponds to the target infeasibility state, where performance requirements conflict with physical capabilities. The time-varying feasible region and constraint conflicts of the two-sided constraints are as follows: Figure 3 As shown, the lower bound of performance requirements increases with changes in operating mode, while the upper bound of physical capabilities decreases as the available margin decreases. The interval between the two is the feasible region at the current moment. When the operating state crosses the feasibility boundary and the lower bound is higher than the upper bound, it enters the infeasible region.

[0143] VI. Constrained Closed-Loop Controller

[0144] The constrained closed-loop controller consists of damping demand feedforward, damping ratio feedback correction, two-sided constraint projection, asymmetric rate of change limitation, and inverse calculation type anti-integral saturation. The damping ratio tracking error and local control gain are shown in Equations (47) and (48), respectively.

[0145] (47)

[0146] (48)

[0147] In the formula, This refers to the tracking error of the regional damping ratio. The local control gain from the wind resistance damping support coefficient to the regional damping ratio is composed of online estimates.

[0148] The damping demand feedforward measure is taken as the lower bound of the performance demand, as shown in Equation (49); the integral state is shown in Equation (50); the nominal damping control command before constraint projection is shown in Equation (51).

[0149] (49)

[0150] (50)

[0151] (51)

[0152] In the formula, This is the damping demand feedforward obtained from model inversion; This represents the integral state of the damping ratio controller; This is the proportional feedback coefficient; The integral feedback coefficient; To constrain the nominal damping control command before projection, the lower bound of the performance requirement is used as both the feedforward and subsequent projection lower bound. This allows the controller to start from the minimum support required to meet the target and make fine adjustments only under the drive of the damping ratio error, avoiding the integrator accumulating from zero.

[0153] The nominal command is projected by the two-sided constraint to obtain the control target, as shown in equations (52) and (53).

[0154] (52)

[0155] (53)

[0156] In the formula, The damping control target is the result of the projection of the two-sided constraint. , These are the lower and upper bounds of the projection interval. As a preferred implementation, when... In cases of constraint intersection, conventional projection is no longer performed; instead, the saturated support target is directly provided by the supervision component. This aligns with the principle of prioritizing the upper bound of physical security; the projection operator also naturally outputs the upper bound value in this degenerate case.

[0157] The actual damping support coefficient is obtained by limiting the projected control target to an asymmetric rate of change, as shown in equations (54) and (55).

[0158] (54)

[0159] (55)

[0160] In the formula, For the first The actual damping support coefficient for each control update cycle; To control the update cycle; , These represent the maximum increase rate and maximum release rate of the damping support coefficient, respectively. A faster input rate is used when damping is insufficient, and a slower release rate is used after stabilization and recovery, in order to suppress frequent adjustments caused by boundary switching and estimation noise.

[0161] To suppress integral saturation during the constraint period, the controller adopts an inverse calculation type anti-integral saturation, as shown in equation (56).

[0162] (56)

[0163] In the formula, This is the anti-integral saturation coefficient for inverse calculation type; The difference between the actual executed quantity and the nominal control quantity, multiplied by the local control gain. The dimension is the same as the tracking error. When the controller is constrained, this back-calculation term causes the integral state to revert, making the internal state of the controller consistent with the actual execution quantity. As a preferred implementation, when entering an infeasible saturation state or performing a safe exit, the integral state can also be directly reset according to the static back-calculation relationship of equation (59); when the confidence flag is estimated... When the integral is frozen, the damping support coefficient is maintained or released; upon re-admission, the integral state can be set to zero or the initial value calculated in reverse according to the current execution amount.

[0164] As a preferred implementation, a quasi-static small-signal analysis can be performed on the closed-loop error: the damping ratio estimation and update stage, along with the execution channel of the damping support command, can be approximated as a first-order lag stage, with an equivalent lag time constant of... The feedforward quantity is refreshed according to the estimated update cycle and maintained between two adjacent updates, so that the feedforward quantity can be regarded as a slowly varying known quantity within the analysis bandwidth; the effects of saturation, amplitude limiting and rate of change limitation are ignored. Thus, the homogeneous part of the closed-loop error dynamics can be organized into the form shown in equation (57).

[0165] (57)

[0166] In the formula, The actual object gain and the local control gain used by the controller The ratio; To estimate the equivalent lag time constant of the execution channel. As can be seen from this characteristic equation, when That is, when the control direction is correct, the coefficients of both the linear term and the constant term are positive, and the error is asymptotically stable; when When the control direction reverses, the constant term becomes negative, the characteristic equation must have a real root in the right half-plane, and the error dynamics exhibit a positive feedback divergence trend. This indicates that correct control direction is a prerequisite for closed-loop stability and forms the design basis for online control direction verification, inversion blocking, and safe exit. The magnitude deviation of the estimated gain only affects dynamic quality; the direction sign is the critical boundary for stability. It should be noted that the processing of refreshing and maintaining the feedforward quantity according to the update cycle ensures that the feedforward channel does not introduce additional feedback within the analysis bandwidth. If, in the specific implementation, the feedforward quantity is instantaneously recalculated from the current estimation result, then the feedforward channel itself has a unit proportional effect on the error, and the equivalent proportionality coefficient is determined by... Become At this point, the stability of the error dynamics also depends on The above conclusion remains unchanged regardless of the symbol.

[0167] VII. Infeasibility, Control Direction Anomalies, and System-Level Handling

[0168] Assuming the direction confidence flag indicates that the direction is reliable, when the bilateral constraints intersect, i.e. At the same time, it is not allowed to exceed the physical capacity of the equipment in order to meet the damping performance. This method clearly stipulates that the upper limit of physical safety takes precedence over the lower limit of performance, and the damping support coefficient actually implemented is taken as the upper limit of physical safety according to formula (58).

[0169] (58)

[0170] After entering the saturated support state, the internal integral state of the controller is calculated in reverse according to formula (59) to make the internal state consistent with the actual execution quantity, so as to avoid the integral state from continuing to accumulate unrealizable damping requirements.

[0171] (59)

[0172] The maximum regional damping ratio that can be achieved when all the resources of the grid-connected wind turbine group are put into use is calculated according to formula (60); the remaining damping deficit and the remaining damping ratio gap are quantified according to formula (61) and formula (62) respectively; the system-level stability support request vector is shown in formula (63).

[0173] (60)

[0174] (61)

[0175] (62)

[0176] (63)

[0177] In the formula, This represents the maximum regional damping ratio achievable when all current grid-connected wind power resources are in operation. The equivalent damping deficit in the region that still needs to be borne by other stable control resources after all the wind turbines in the grid are put into operation; This represents the remaining damping ratio gap; This is the system-level stability support request vector. This vector transmits the constraint conflict state, remaining damping deficit, and remaining damping ratio gap to the system-level stability control layer as quantitative inputs for coordinating other damping resources. The supervisory control state transition for regional damping supports is as follows: Figure 4 As shown: When the feasibility margin is sufficient, the system operates in a normal and feasible projected closed-loop state; when the margin decreases to below the safety threshold, it enters a critically feasible state, the controller suppresses integration and outputs a warning; when the constraints continue to tighten and cause a crossover, it enters a target infeasible state, executes according to the physical upper limit and outputs the remaining damping deficit; when the control direction is abnormal or the estimation is unreliable, it switches to a direction or measurement abnormal state and performs actions such as blocking inversion, safety easing and freezing or inverse integration; after the abnormal factors are eliminated and reconfirmed, it returns to the normal closed-loop state.

[0178] When the control direction is reliable flag When the gain inversion based on the control effectiveness gain is stopped, as shown in Equation (64), the additional damping channel in the region is smoothly exited according to the safe release rate, as shown in Equation (65).

[0179] (64)

[0180] (65)

[0181] In the formula, This is the safe release rate set in case of abnormal control direction. After the damping support coefficient drops to zero, only the regional additional damping channel is withdrawn; the original grid control of the wind turbine cluster is not shut down. After safe withdrawal, the control center recalculates the cluster control effectiveness gain, reconstructs the effective cluster set and assigns weights; only after continuous confirmation of the direction on both the model side and the data side is the regional damping inversion control allowed to be reactivated.

[0182] As a preferred implementation method, estimate the trust flag. The handling of abnormal situations is coordinated with the direction: the integral link is frozen, and the damping support coefficient is maintained or released at a slow release rate, and gain inversion is not performed; when the confidence flag is restored and the continuous confirmation condition is met, the closed-loop control is restored according to the re-access procedure, and the slow release rate can be the maximum release rate or a certain proportion thereof.

[0183] VIII. Energy Dissipation Mechanism of Regional Oscillations

[0184] The equivalent energy and energy dissipation relationship of the regional dominant oscillation are shown in equations (66) to (68).

[0185] (66)

[0186] (67)

[0187] (68)

[0188] In the formula, The first term is the equivalent kinetic energy of the relative motion between the two coherent regions, and the second term is the equivalent synchronous potential energy corresponding to the relative work angle shift. For the additional regional oscillation energy dissipation power formed by the grid-connected wind turbine group, when , The dissipated power is non-negative. This result shows that the proposed method does not merely change an abstract damping parameter, but rather continuously dissipates the dominant oscillation energy of the region through an additional active power channel that is opposite to the relative oscillation velocity of the region.

[0189] IX. Specific Implementation Process

[0190] The implementation process of this method may include the following steps.

[0191] S1. Regional Dynamic Measurement Acquisition. Measurement information such as key bus voltage phase angle, node frequency, tie-line active power, active power output of the grid-connected wind turbine group, bidirectional active power regulation margin, and converter limiting status are acquired through PMU / WAMS. After time synchronization, outlier removal, and filtering, the measurement data enters the regional modal analysis stage.

[0192] S2. Dominant Mode and Coherent Region Identification. The multi-channel overall least squares matrix pencil method is used to obtain the eigenvalues ​​and damping ratio estimates of the dominant modes online. Based on the phase and inertia center frequency response of the dominant modes, the system is divided into two main coherent regions, and the equivalent relative work angle and relative oscillation angular velocity of the regions are calculated.

[0193] S3. Update the equivalent parameters of the region. The equivalent inertia, intrinsic damping and synchronous stiffness of the region are updated near the current operating point by model projection, parameter inversion or recursive estimation. The intrinsic damping can be inversely solved by the closed-loop damping ratio according to Equation (33) to provide input for the performance requirement inversion.

[0194] S4. Calculation of control effectiveness. Based on the state matrix of the current operating point, the additional active power input matrix of each group, the left and right characteristic vectors of the dominant mode and the relative velocity observation vector, calculate the control effectiveness gain of a single group, construct an effective group set, and calculate the allocation weight and overall control effectiveness gain according to equations (24) to (26).

[0195] S5. Online verification of control direction. The data-side gain is estimated online using recursive least squares, and the sign and amplitude are verified against the model-side results. After continuous confirmation, the confidence flag of the control direction is set to 1.

[0196] S6. Generation of bilateral constraints. The lower bound of performance requirements is calculated based on the regional damping reference value, equivalent parameter estimation, and intrinsic damping estimation; the upper bound of physical capabilities is calculated based on the bidirectional active power regulation margin of each generator group, the amplitude of slow oscillation, and the hardware upper limit.

[0197] S7. Feasibility Monitoring and Mode Selection. Calculate the feasibility margin of both-sided constraints, and combine the control direction confidence indicator and the estimated confidence indicator to select the operating state, such as normal closed loop, critically feasible, objective infeasible, or safe exit.

[0198] S8. Constrained Closed-Loop Control. Under feasible conditions, damping demand feedforward, damping ratio feedback, gain inversion, two-sided constraint projection, asymmetric rate of change limitation, and inverse calculation anti-integral saturation are executed sequentially to generate the actual damping support coefficient.

[0199] S9. Multi-unit group execution. Each grid wind turbine group generates additional damped active power instructions according to the allocated weight, which are then superimposed on the original grid active power reference after local limiting, as shown in equations (69) to (71).

[0200] (69)

[0201] (70)

[0202] (71)

[0203] In the formula, For the first Unlimited additional damped active power commands for individual generator groups; The actual additional damping active power after local saturation; The active power reference value for the generator group after adding regional damping support; This is the reference value for the original network active power.

[0204] S10, System-level handling. When constraints conflict, the damping support coefficient is limited to a physical upper limit, and the remaining damping deficit and remaining damping ratio gap are output; when the control direction is abnormal or the estimation is unreliable, the gain inversion is blocked and the additional damping channel in the region is smoothly exited at a safe release rate, and re-entry is allowed after the conditions are restored.

[0205] 10. Calculation Examples

[0206] To illustrate the role and numerical consistency of each step in the method, this section presents a set of design-oriented numerical examples focusing on control mechanisms and constraint logic. These examples cover scenarios such as control effectiveness gain, damping ratio closed-loop adjustment, feasibility determination of two-sided constraints, residual damping deficit output, and safe exit from control direction anomalies. These examples are for design-oriented numerical verification and do not constitute limitations on specific system parameters. In engineering implementation, it is advisable to verify the control parameters in conjunction with electromechanical or electromagnetic transient models of the actual regional power grid.

[0207] The parameters for the example are shown in Table 2.

[0208] Table 2 Parameters for Design Examples

[0209]

[0210] Based on the parameters in Table 2, the undamped natural angular velocity, synchronous stiffness and target damping requirements of the region are calculated according to equations (72) to (74).

[0211] (72)

[0212] (73)

[0213] (74)

[0214] In this case, the undamped natural angular frequency under weakly damped conditions is approximately equal to the damped oscillation angular frequency, therefore, we take... .

[0215] The intrinsic damping of the region under the initial working condition and after the deterioration of the operating mode is 5.23 pu and 1.96 pu, respectively. The corresponding minimum damping support coefficients are calculated according to Equation (75) and Equation (76), respectively.

[0216] (75)

[0217] (76)

[0218] To facilitate a focused examination of the control logic, this example uses a constant control effectiveness gain. This is used to simulate the gain calculated online during actual operation. In actual implementation, this gain is calculated online using the small-signal model and measured data at the current operating point, as described above, and is refreshed as the operating mode changes. The timing sequence of the operating conditions is shown in Table 3.

[0219] Table 3 shows the time sequence of the calculation example.

[0220]

[0221] The closed-loop response of the region-dominant mode damping ratio is as follows: Figure 5As shown. Under the initial weakly damped condition, the controller engages the regional damping support based on the minimum damping support coefficient, and the online estimated damping ratio gradually approaches the reference value of 0.18; After the intrinsic damping decreases, the controller increases the damping support coefficient and restores it to near the target. After s, the target becomes infeasible due to the decrease of the physical upper bound, and the damping ratio is no longer forcibly maintained at the reference value, but runs near the level corresponding to the physical upper bound; After an anomaly in the s-direction, the additional damping channel in the region exits at a safe rate. This response reflects the control principles of closed-loop tracking when performance is achievable, adherence to physical boundaries when performance is unachievable, and prohibition of inversion when the direction is abnormal. It should be noted that this response process is an illustrative result of a design example, and the actual dynamics are affected by the modal estimation window and update cycle.

[0222] The two-sided constraint boundary and the actual damping control quantity are as follows Figure 6 As shown, the lower bound of performance requirements and the upper bound of physical capabilities are displayed in separate, independent coordinate regions. After s, the lower bound of the performance requirement is approximately 11.95 pu, while the upper bound of the physical capability drops to 9.5 pu, resulting in a conflict between the two constraints; the actual damping support coefficient is limited to the physical upper bound instead of continuing to follow the unachievable lower bound of the performance.

[0223] The response of the relative oscillation velocity of the region under different damping control methods is as follows: Figure 7 As shown, the oscillation decays slowest when there is no damping support; the performance of fixed damping support deteriorates after the operating mode changes; this method can increase the support strength according to the damping requirements in the feasible stage of double-sided constraints, and the oscillation decays faster; after entering the infeasible stage, the controller still provides the maximum positive damping allowed by the current physical capabilities, but no longer takes the achievement of the target damping as an achievable conclusion.

[0224] Infeasible deficit output and control direction supervision, such as Figure 8 As shown. After the constraint conflict occurs, the maximum additional regional equivalent damping that the wind turbine group can provide under the physical upper limit is calculated according to formula (77), the corresponding regional total equivalent damping is calculated according to formula (78), and the remaining damping deficit is calculated according to formula (79), which can be directly used as the demand for other damping resources at the system level; the maximum damping ratio that the wind turbine group can achieve under the physical upper limit is calculated according to formula (80), and the remaining damping ratio gap is calculated according to formula (81). Figure 8 The lower part displays the control direction confidence flag. s changes from 1 to 0, after which gain inversion is disabled and a safe exit is triggered.

[0225] (77)

[0226] (78)

[0227] (79)

[0228] (80)

[0229] (81)

[0230] Regional dominant oscillation energy decay envelope such as Figure 9 As shown. This method provides higher positive dissipation intensity during the feasible operation phase; the fixed damping capacity decreases after the operation mode changes because the support strength is no longer matched; the energy attenuation result is consistent with the positive dissipation mechanism of equations (67) and (68).

[0231] Table 4 compares the undamped support, fixed damped support, and this method from the perspective of capability.

[0232] Table 4 Comparison of Scheme Capabilities

[0233]

[0234] In summary, compared with existing schemes that rely on empirical tuning or approximate fixed gain for additional damping control, this embodiment constructs an online control effectiveness gain, making the strength of the influence of the grid-connected wind turbine group on the regional dominant oscillation channel calculable and comparable, and establishing the determination of damping support strength based on online quantitative evidence. Through dual-source direction verification on both the model and data sides, the stability prerequisite of correct control direction is transformed into a closed-loop entry condition. When the direction is unreliable, gain inversion is blocked, and the additional damping channel is smoothly exited, avoiding the risk of negative damping and positive feedback caused by incorrect direction judgment. Furthermore, the exit process does not affect the original grid-connected control of the wind turbine group. After safe exit, the gain is recalculated, the effective turbine group set is reconstructed, and weights are allocated. Re-entry is possible after the conditions are restored, ensuring the safety and recoverability of the control.

[0235] By constructing a time-varying two-sided feasible region with a lower bound for performance requirements and an upper bound for physical capabilities, the target damping and equipment capabilities are matched online. Feasibility margin supervision distinguishes between normal, critical, and infeasible states, enabling the controller to track the target in a closed loop when performance is achievable and to execute according to the physical upper bound when capabilities are limited, thus avoiding the continuous accumulation of unachievable damping requirements. Feedforward, feedback, projection, asymmetric speed limiting, and inverse calculation-type anti-integral saturation work together to achieve closed-loop tracking of the region-dominant mode damping ratio within the feasible region and suppress integral saturation and frequent boundary switching during the constraint period.

[0236] When constraints conflict, the system outputs the remaining damping deficit and the remaining damping ratio gap, providing quantitative input for the system-level stability control layer to coordinate other damping resources, thus forming a handling mechanism that connects wind power-side regulation with system-side coordination. Furthermore, the online estimation of the regional damping ratio provides clear algorithm selection, window parameters, and reliability judgments. The regional slow loop and the local fast loop of the wind turbine cluster form a time-scale separation, facilitating integration with existing wide-area measurement systems and site control systems. Therefore, the grid-connected wind turbine cluster transforms from a fixed additional damping resource into a regional damping support resource with calculable control gain, discriminative control direction, constrained closed-loop regulation, identification of infeasible states, and output of system stability requirements.

[0237] This embodiment also provides a regional damping support control device for a grid-connected wind turbine cluster in a high-proportion renewable energy power grid, including an estimation module, a gain construction module, a constraint generation module, a monitoring control module, and an execution module. The estimation module is used to acquire online estimates of the regional dominant mode damping ratio and the regional relative oscillation angular velocity; the gain construction module is used to construct online the control effectiveness gain from the grid-connected wind turbine cluster to the regional dominant oscillation channel, and to perform online verification of the control direction indicated by the control effectiveness gain, including comparing the control effectiveness gain with the data-side gain; the constraint generation module is used to determine the lower bound of the performance requirement based on the difference between the regional total equivalent damping requirement corresponding to the target damping ratio and the online estimated regional intrinsic damping, through the control effectiveness gain inversion, and to determine the upper bound of the physical capability based on the bidirectional symmetrical active power margin of the grid-connected wind turbine cluster and the amplitude of the regional relative oscillation angular velocity; the monitoring module... The control module stops the inversion based on control effectiveness gain and exits the regional additional damping channel when the control direction verification fails. When the control direction verification passes and the lower bound of performance requirement does not exceed the upper bound of physical capacity, it adjusts the damping support coefficient in a closed loop within the constraint interval formed by the two to make the estimated damping ratio track the target damping ratio. When the control direction verification passes and the lower bound of performance requirement exceeds the upper bound of physical capacity, it limits the damping support coefficient to the upper bound of physical capacity and calculates the remaining damping deficit, outputting it to the system-level stability control layer. The execution module generates additional active power in the opposite direction to the regional relative oscillation angular velocity based on the damping support coefficient and superimposes it on the original active power reference value of the grid-connected wind turbine group. Each module can be implemented by the software or hardware functions of the same controller, and its specific implementation method is the same as the aforementioned method.

[0238] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for regional damping support control of wind turbine clusters in a high-proportion renewable energy power grid, characterized in that, include: Online acquisition of estimated damping ratio of dominant mode in the region and relative oscillation angular velocity of the region; The control effectiveness gain of the online construction of the grid-connected wind turbine group to the regional dominant oscillation channel, and its product with the damping support coefficient constitutes the additional equivalent damping provided by the grid-connected wind turbine group; The control direction of the control effectiveness gain is verified online, including comparing the control effectiveness gain with the data-side gain. The lower bound of the performance requirement is obtained by inverting the control effectiveness gain based on the difference between the total equivalent damping requirement of the region corresponding to the target damping ratio and the intrinsic damping of the region estimated online; the upper bound of the physical capability is determined based on the bidirectional symmetrical active power margin of the grid-connected wind turbine group and the amplitude of the relative oscillation angular velocity of the region. When the verification is passed and the lower bound of the performance requirement does not exceed the upper bound of the physical capability, the damping support coefficient is adjusted in a closed loop within the constraint range formed by the two to make the estimated damping ratio track the target damping ratio. When the verification is passed and the lower bound of the performance requirement exceeds the upper bound of the physical capacity, the damping support coefficient is limited to the upper bound of the physical capacity, the remaining damping deficit is calculated and output to the system-level stability control layer. If the verification fails, stop the inversion based on the control effectiveness gain and exit the regional additional damping channel; An additional active power, which is opposite to the regional relative oscillation angular velocity, is generated based on the damping support coefficient and superimposed on the original active power reference value of the grid-connected wind turbine group.

2. The method for regional damping support control of wind turbine clusters in high-proportion new energy power grids according to claim 1, characterized in that: The lower bound of the performance requirement is determined as follows: Calculate the total equivalent damping requirement of the region based on the target damping ratio, the estimated equivalent inertia of the region and the estimated equivalent synchronous stiffness of the region. Divide the difference between the total equivalent damping requirement of the region and the intrinsic damping of the region by the control effectiveness gain, and take the larger value between the result and zero as the lower bound of the performance requirement; The regional intrinsic damping is obtained online by inverse solving based on the mapping relationship between the regional dominant mode damping ratio and the regional total equivalent damping, using the damping ratio estimate, regional equivalent inertia estimate, regional equivalent synchronous stiffness estimate, control effectiveness gain, and current damping support coefficient.

3. The method for regional damping support control of wind turbine clusters in high-proportion new energy power grids according to claim 1, characterized in that: The control effectiveness gain from the online construction of the grid-connected wind turbine group to the regional dominant oscillation channel includes: Based on the linearized state matrix of the current operating point of the regional power grid, the weighted equivalent input matrix of the grid-connected wind turbine group, the right and left eigenvectors of the regional dominant oscillation mode, and the observation vector characterizing the regional relative oscillation angular velocity, the sensitivity of the regional dominant oscillation mode eigenvalues ​​to the damping support coefficient is calculated. The sensitivity is transformed by damping ratio and then multiplied by twice the square root of the product of the regional equivalent inertia and the regional equivalent synchronous stiffness to obtain the control effectiveness gain.

4. The method for regional damping support control of wind turbine clusters in high-proportion new energy power grids according to claim 3, characterized in that: The single-unit control effectiveness gain of each candidate grid-connected wind turbine group is calculated separately. The grid-connected wind turbine groups whose single-unit control effectiveness gain is greater than the set admission threshold constitute the effective turbine group set. Based on the bidirectional symmetrical active power margin and single-unit control effectiveness gain of each grid-connected wind turbine group in the effective turbine group set, the comprehensive support capability index of each grid-connected wind turbine group is calculated, the power allocation weight of each grid-connected wind turbine group is obtained by normalization, and the control effectiveness gain is obtained by weighted aggregation according to the power allocation weight.

5. The method for regional damping support control of wind turbine clusters in high-proportion new energy power grids according to claim 1, characterized in that: The online verification includes: The control effectiveness gain on the model side is calculated from the linearized model of the current operating point of the regional power grid. A local autoregressive model is established using a data sequence consisting of the estimated damping ratio and the increment of the damping support coefficient. The data-side gain is obtained through online identification using recursive least squares. Within a continuously set number of update cycles, if both the model-side control effectiveness gain and the data-side gain are positive, and the relative deviation between the data-side gain and the local control gain determined by the model-side control effectiveness gain, the regional equivalent inertia, and the regional equivalent synchronization stiffness does not exceed the set tolerance, the control direction is determined to be reliable; otherwise, the control direction is determined to be unreliable.

6. The method for regional damping support control of wind turbine clusters in high-proportion new energy power grids according to claim 4, characterized in that: The upper bound of the physical capability is determined in the following manner: The bidirectional symmetrical active power margin of each grid-connected wind turbine group is taken as the minimum value among the active power margin of the grid-connected wind turbine group that is adjusted upward, the active power margin that is adjusted downward, and the margins corresponding to the converter current, power and thermal stability limits. The square of the relative oscillation angular velocity of the region is subjected to a first-order recursive filter and the square root is taken to obtain the slow-varying root mean square amplitude. The larger value between the slow-varying root mean square amplitude and the minimum amplitude threshold is taken as the effective oscillation amplitude. Based on the bidirectional symmetrical active power margin, power distribution weight, and effective oscillation amplitude of each grid-connected wind turbine group within the effective turbine group set, the upper limit of the damping support coefficient of each grid-connected wind turbine group is calculated. The minimum value is compared with the hardware upper limit of the damping coefficient, and the smaller value is taken as the upper limit of the physical capability.

7. The method for regional damping support control of wind turbine clusters in high-proportion new energy power grids according to claim 1, characterized in that: The closed-loop adjustable damping support coefficient includes: The lower bound of the performance requirement is used as the damping requirement feedforward. The difference between the target damping ratio and the estimated damping ratio is calculated as the tracking error; The tracking error is subjected to proportional-integral (PI) calculation, and the result of the PI calculation is divided by the local control gain and then superimposed on the damping demand feedforward to obtain the nominal damping control command; the local control gain is determined by the control effectiveness gain, the regional equivalent inertia, and the regional equivalent synchronization stiffness. The nominal damping control command is projected onto the interval formed by the lower bound of the performance requirement and the upper bound of the physical capability, and an asymmetric rate of change constraint is applied to the projection result, wherein the rate of increase of the damping support coefficient is greater than the release rate, to obtain the actual damping support coefficient. The difference between the actual damping support coefficient and the nominal damping control command is converted into a local control gain and then fed back to the integral stage of the proportional-integral operation.

8. The method for regional damping support control of wind turbine clusters in high-proportion new energy power grids according to claim 7, characterized in that: When the lower bound of performance requirements exceeds the upper bound of physical capabilities, the actual damping support coefficient is taken as the upper bound of physical capabilities, and the integral state is calculated in reverse based on the actual damping support coefficient, damping requirement feedforward, and tracking error. The remaining damping deficit is the larger of the following: the total equivalent damping demand of the region minus the intrinsic damping of the region, minus the product of the control effectiveness gain and the upper bound of the physical capability, and zero. Calculate the maximum damping ratio that can be achieved after the upper limit of physical capacity is put into operation, and take the larger value between the difference between the target damping ratio and the maximum damping ratio and zero as the remaining damping ratio gap; The system-level stability support request information, consisting of the infeasibility state flag, the remaining damping deficit, and the remaining damping ratio gap, is output to the system-level stability control layer.

9. The method for regional damping support control of wind turbine clusters in high-proportion new energy power grids according to claim 4, characterized in that: The additional active power is decomposed according to the power allocation weight of each grid-connected wind turbine group. Each grid-connected wind turbine group generates an additional damped active power command that is opposite to the regional relative oscillation angular velocity based on the allocation result. After local power limiting, it is superimposed on the original active power reference value of the grid-connected wind turbine group.

10. A regional damping support control device for wind turbine clusters in a high-proportion renewable energy power grid, used to implement the method as described in claim 1, characterized in that, include: The estimation module is used to obtain online estimates of the damping ratio of the dominant mode in the region and the relative oscillation angular velocity of the region. The gain construction module is used to construct the control effectiveness gain of the grid-connected wind turbine group to the regional dominant oscillation channel online, and to perform online verification of the control direction of the control effectiveness gain, including comparing the control effectiveness gain with the data-side gain. The constraint generation module is used to determine the lower bound of performance requirements based on the target damping ratio and the online estimated regional intrinsic damping, through the control effectiveness gain inversion, and to determine the upper bound of physical capacity based on the bidirectional symmetrical active power margin of the grid-connected wind turbine group and the amplitude of the regional relative oscillation angular velocity. The monitoring and control module is used to stop the inversion based on the control effectiveness gain and cause the regional additional damping channel to exit when the online verification fails. When the online verification passes and the lower bound of the performance requirement does not exceed the upper bound of the physical capability, the damping support coefficient is adjusted in a closed loop within the constraint interval formed by the two to make the estimated damping ratio track the target damping ratio; when the online verification passes and the lower bound of the performance requirement exceeds the upper bound of the physical capability, the damping support coefficient is restricted to the upper bound of the physical capability, the remaining damping deficit is calculated and output to the system-level stability control layer. The execution module is used to generate additional active power that is opposite to the regional relative oscillation angular velocity based on the damping support coefficient, and superimpose it on the original active power reference value of the grid wind turbine group.

Images