New energy field group broadband oscillation self-stabilization control method and system based on passivity constraint

Abstract

The application discloses a new energy field group broadband oscillation self-stabilization control method and system based on passivity constraint, establishes small signal admittance matrix of each converter port and network node admittance matrix under multiple operating conditions, and retains multi-port coupling through node elimination; the minimum eigenvalue of the normalized admittance Hermite part of the network and the device port is calculated to form a port passivity index, and the minimum singular value of the return difference matrix is used to screen a risk frequency band and determine a stability symbol; the leading device and candidate parameters are determined in combination with critical roots, critical eigenvectors and normalized parameter sensitivity; parameter increments are solved under the constraint, are issued in a projection, backtracking and retesting mode, and are rolled back to the last set of safe parameters when strict stability or operating constraint is not satisfied. The application can form an implementable closed loop from cross-condition passivity margin, leading device and parameter positioning, to constrained parameter solving and safe online updating under the premise of retaining the port coupling relationship of the heterogeneous field group.

Description

Technical Field

[0001] This invention belongs to the field of new energy grid connection and power system stability control technology, and relates to a method and system for broadband oscillation self-stabilization control of new energy power field groups based on passive constraints. Background Technology

[0002] With the large-scale integration of new energy sources such as wind and solar power, the power system is gradually shifting from a synchronous machine-dominated structure to a power electronic converter-dominated structure. In new energy power plants, grid-connected and grid-connected converters coexist, and their control mechanisms differ significantly. This leads to strong coupling, multi-frequency band, and nonlinear characteristics in the system impedance, easily causing broadband oscillations such as subsynchronous, mid-frequency, and high-frequency oscillations, affecting system stability. Existing oscillation suppression methods typically target single frequency bands or local devices, lacking a unified quantitative description of stability across the entire frequency domain. Furthermore, control parameters often rely on empirical tuning, making it difficult to effectively integrate stability and control design. Passive theory can uniformly characterize system stability from the perspective of impedance positive reality, but current research largely focuses on stability determination and has not yet developed a multi-frequency band integrated self-stabilizing control method based on passive constraints. Therefore, it is necessary to propose a broadband oscillation self-stabilizing control method for new energy power plants based on passive theory.

[0003] Existing solutions typically treat the field group as equivalent to a single source-side impedance and load-side impedance, or adjust the gain of a controller only based on the oscillation frequency. It is difficult to simultaneously retain multi-node port coupling, unify the stability margin scale of devices with different rated capacities, and directly constrain the device positioning results to the control parameters that can be issued. Summary of the Invention

[0004] The purpose of this invention is to provide a broadband oscillation self-stabilizing control method and system for new energy field groups based on passive constraints. This method can form an implementable closed loop, from cross-condition passive margin, dominant equipment and parameter positioning, to constrained parameter solving and safe online updating, while preserving the port coupling relationship of heterogeneous field groups.

[0005] To achieve the above objectives, the solution of the present invention is:

[0006] A method for broadband oscillation self-stabilization control of new energy field clusters based on passive constraints includes the following steps:

[0007] Step 1, in the set of operating conditions Operating conditions in Below, using the steady-state operating points of the grid-type converter and the mesh-type converter as linearization points, the reduced-order impedance matrix of the field group multi-port is obtained. ;

[0008] Step 2: Divide the frequency domain into multiple characteristic frequency bands. Define passive indicators for device ports and passive indicators of network ports Assemble the network admittance YN and the device admittance YD and construct the return difference matrix. ,by Screening for risk bands;

[0009] Step 3, when the port passive indicator is below the threshold, When the value falls below the risk threshold or the return complex eigenvalue enters the right half-plane, the critical frequency, critical operating condition, and critical eigenvector are determined; the real part of the rightmost complex root and the normalized sensitivity of the port passive index to the candidate control parameters are calculated, and the dominant converter and the set of candidate control parameters are determined accordingly.

[0010] Step 4, increment the candidate control parameters As the decision variable, under the constraints that the passive index of the target port is not lower than the corresponding threshold, the real part of the rightmost complex root of the return difference is not greater than the preset negative damping limit, the parameters are within the allowable range, the controller is internally stable, the voltage, current and modulation amount do not exceed the limits, and the control delay is limited, the control parameter increment that minimizes the change of the normalized parameter is solved.

[0011] Step 5: Map the obtained control parameter increment to at least one of the virtual impedance, phase-locked loop damping, feedforward compensation, virtual inertia, virtual damping, or droop control loop of the dominant converter, and determine the actual impedance increment based on the difference between the reduced-order impedance matrix before and after the update.

[0012] Step 6: Issue control parameters according to the update rules with projection and backtracking step size, remeasure or identify port admittance and recalculate port passive indicators, return difference eigenvalues ​​and operating constraints; roll back to the previous set of safety parameters when strict stability check fails, target port passive indicators deteriorate beyond the allowable amount or operating quantity exceeds the limit; stop updating when the complex root damping and operating constraints are met for a preset number of consecutive times.

[0013] The specific content of step 1 is as follows:

[0014] In the set of operating conditions Operating conditions in Below, using the steady-state operating points of the grid-type converter and the mesh-type converter as linearization points, we obtain the first... Dynamic state of the converter Port voltage disturbances and port current disturbances are denoted as follows: , and ;

[0015] Establish the small-signal admittance matrix of each converter port In addition to the nodal admittance matrices of the collector lines, transformers, and external power grid, internal nodes are eliminated to obtain a reduced-order impedance matrix that preserves the coupling relationship between the observation ports. .

[0016] In step 1, the first The port voltage disturbance and port current disturbance of the converter are represented by two-dimensional vectors in a synchronously rotating dq coordinate system, wherein the port small-signal admittance matrix is... It is obtained by linearizing the converter power stage, filter, digital control delay and control loop at the steady-state operating point; when using positive and negative order coordinates, the dq admittance matrix is ​​converted into an order domain admittance matrix containing positive and negative order coupling terms through a preset coordinate transformation matrix.

[0017] In step 1, the network node admittance matrix formed by the collector lines, transformers, and external power grid is arranged according to the set of reserved port nodes. and the set of internal nodes of the element to be eliminated Divided into blocks , , and ;in, This represents the self-admittance and mutual admittance submatrices between the retained port nodes. This represents the coupling submatrix of the internal node voltage to the injected current at the reserved port. This represents the coupling submatrix that preserves the port voltage's relationship to the injected current at the internal nodes. Represents the self-admittance and mutual admittance submatrices between internal nodes; in When reversible, according to Obtain the network reserved port admittance, and assemble the small-signal admittances of each converter port into... and in accordance with and Obtain the reduced-order impedance matrix of the field group multi-port.

[0018] In step 2, the first is defined. Normalized port admittance of the converter Its Hermitian part is and according to Calculate the first The converter in the first Passive indicators of device ports in each characteristic frequency band Construct an open-loop matrix The sum returns the difference matrix. and calculate As an indicator of risk frequency band distance, the stable symbol is... The characteristic trajectory is relatively Number of points around the point or The rightmost complex root is determined.

[0019] In step 3, at the critical frequency and critical operating conditions, the sensitivity of the passive index of the port is calculated using the critical eigenvector, and the sensitivity of the real part of the rightmost complex root to the candidate parameters is calculated using the central finite difference method. The responsibility is formed according to the sensitivity, the allowable adjustment of the parameters, the residual current of the equipment, and the modulation margin.

[0020] In step 4, with The objective function is a weighted sum of slack variables, with the objective port passive index not lower than a set value after constraint update, and the real part of the rightmost complex root of the return difference not greater than a certain value. , To preset the negative damping limit and satisfy parameter boundaries, internal stability, current, voltage, modulation ratio and additional control output limits.

[0021] In step 5, at least one of the following parameters is adjusted for the grid-type converter: virtual impedance parameter, phase-locked loop damping compensation parameter, voltage feedforward parameter, current feedforward parameter, or power feedforward parameter; at least one of the following parameters is adjusted for the grid-type converter: virtual inertia, virtual damping, droop coefficient, or additional virtual impedance parameter; the parameter adjustment shall not change the internal stability of the controller and shall be constrained by the converter current and modulation margin.

[0022] In step 6, the following is adopted: Update control parameters; if the root damping does not reach the threshold, the passive performance of the port deteriorates beyond the allowable limit, the voltage, current, or modulation exceeds the limit, communication is abnormal, or the identification confidence is insufficient, cancel the update and reduce the value. .

[0023] A broadband oscillation self-stabilizing control system for a new energy field cluster based on passive constraints includes a processor, a memory, an impedance measurement or identification interface, an operating condition acquisition interface, and a control parameter distribution interface connected to a grid-connected converter and a network-connected converter; the memory stores instructions, which, when executed by the processor, implement the method described above.

[0024] After adopting the above scheme, the core idea of ​​this invention is to retain the multi-port coupling relationship of heterogeneous field groups, characterize the local energy characteristics with passive indicators of equipment and network ports, screen risk frequency bands with the minimum singular value of return difference, and then determine the stable sign by the number of generalized Nyquist loop points or the complex eigenvalue of return difference; on this basis, the dominant device is located by using critical roots and parameter sensitivity, and broadband oscillation self-stabilizing control is achieved through constrained parameter solving, supervision and verification, backtracking and rollback of failure.

[0025] The present invention has the following beneficial effects:

[0026] (1) Compared with the scheme that simplifies the field group to a single scalar impedance, the present invention preserves the multi-port coupling relationship between the heterogeneous converter, the collector network and the external power grid through node admittance assembly and Kron elimination, so that the stability evaluation object corresponds to the actual field group topology.

[0027] (2) Quantify the local energy characteristics by passive indicators of equipment and network ports, screen risk frequency bands by the minimum singular value of return difference, and then strictly determine the stable sign by the number of generalized Nyquist winding points or complex characteristic roots of return difference, so as to avoid directly using the overall reduced-order impedance Hermitian index or single amplitude intersection point that does not guarantee monotonicity as sufficient criteria.

[0028] (3) By using critical roots, critical eigenvectors and normalized parameter sensitivity, the risk frequency band is mapped to specific equipment and adjustable parameters, and the allowable adjustment of parameters and operating margin are incorporated into the responsibility, so that sensitivity analysis can directly serve the allocation of control responsibility and optimization dimensionality reduction.

[0029] (4) By solving the joint constraints with the control parameter increment as the decision variable, the parameters that can be written into the PLL, virtual impedance, feedforward, virtual inertia, virtual damping or droop are directly output, avoiding the lack of a unique mapping between the impedance increment and the abstract fitting function.

[0030] (5) By using projection, backtracking, shadow model verification, slope limitation, online retesting and failure rollback constraint parameter updates, the system can restore to the previous set of safety parameters when strict stability, port passive indicators or operation constraints are not met, thus forming a closed-loop safety measure that can be verified and rolled back. Attached Figure Description

[0031] Figure 1 This is a block diagram of the overall structure of the control strategy of the present invention;

[0032] Figure 2 Comparison chart of passive performance parameters of the selected ports for the 124Hz target frequency;

[0033] Figure 3 The graph shows a comparison between the risk band return difference index and the real part of the rightmost complex root, where (a) is the minimum singular value of the return difference and (b) is the real part of the rightmost complex root.

[0034] Figure 4 The results of electromagnetic transient identification are shown in the figure, where (a) is the risk mode envelope, (b) is the comparison between the theoretical and electromagnetic transient growth rates, and (c) is the amplitude ratio before and after the mode envelope.

[0035] Figure 5 The diagram shows the operational constraint verification of the constrained collaborative scheme, where (a) represents the modulation peak value of the grid-following and grid-connected converters, and (b) represents the steady-state active power of each 35kV bus section. Detailed Implementation

[0036] The technical solution and beneficial effects of the present invention will be described in detail below with reference to the accompanying drawings.

[0037] like Figure 1 As shown, this invention provides a broadband oscillation self-stabilizing control method for new energy field clusters based on passive constraints, comprising the following steps:

[0038] Step 1: Unified Impedance Modeling of New Energy Field Clusters

[0039] For each operating condition The operating point obtained from power flow calculations or steady-state measurements is used as the linearization point. The dynamic state, port voltage disturbance, and port current disturbance of the converter are denoted as follows: , and The following linearized state equations and port current output equations are established:

[0040]

[0041] The two equations above together constitute a linearized state-space model at the same operating point. The first equation describes the state perturbation. The second equation describes the port current disturbance as it dynamically changes over time. Consisting of state disturbances and port voltage disturbances A joint decision. In the formula, For the set of operating conditions, For one of the operating conditions, For converter equipment serial number, Let be the state perturbation vector. These are the port voltage and port current disturbance vectors in the synchronously rotating dq coordinate system, respectively, with superscripts indicating their respective perturbations. Indicates transpose. Represents the derivative with respect to time. The state matrix, For port voltage input matrix, For port current output matrix, This is the direct transfer matrix from port voltage to port current. For the first The state dimension of the converter.

[0042] Performing a Laplace transform on the above state equation yields the port small-signal admittance matrix. ;when When reversible, the first The small-signal impedance matrix of the converter port is . Indicates the first The converter changes from port current disturbance to port voltage disturbance. Frequency domain relationships. Among them, Let the complex frequency variable be the Laplace transform. ; For the real part, The imaginary unit, Angular frequency is used; when calculating the frequency response, take... .

[0043] Small-signal admittance of each converter port Assemble into block diagonal device admittance according to reserved port order. The branch admittances of the collector lines, transformers, and external power grid are assembled into a network node admittance matrix according to the actual topology. By reserved port node set and the set of internal nodes of the element to be eliminated right Obtained by dividing into blocks , , and ,in, This represents the self-admittance and mutual admittance submatrices between the retained port nodes. This represents the coupling submatrix of the internal node voltage to the injected current at the reserved port. This represents the coupling submatrix that preserves the port voltage's relationship to the injected current at the internal nodes. Represents the self-admittance and mutual admittance submatrices between internal nodes; when When reversible, Krone elimination is used to obtain the network reserved port admittance. , and then and .therefore, It includes network admittance formed by topology and passive branches. It also includes the device admittance formed by the small-signal models of each converter. ; This is the reduced-order impedance matrix for the field group multiport used in subsequent stability assessments. It only degenerates into a scalar source-load series impedance in the special case of a single-port circuit with negligible coupling.

[0044] Step 2: Port passive indicators, risky frequency band screening, and strict stability assessment

[0045] For the first Define the normalized port admittance of the converter. Its Hermitian part is The frequency domain is divided into multiple characteristic frequency bands. , for the Each frequency band defines the passive parameters of the device port. Accordingly, the normalized network port admittance is defined. and its Hermite part Network port passive indicators are based on calculate. or The larger the value, the higher the dissipation margin of the corresponding port in that frequency band; its threshold is determined by the upper bound of the identification or fitting error and the engineering safety margin.

[0046] Equipment Admittance The network admittance is obtained by assembling the linearized models or port identification results of each converter in Step 1 according to the port order. The network node admittance matrix, established from known line, transformer, and external power grid parameters, is obtained through internal node elimination; both are calculated or identified based on the current operating conditions and verification frequency, and are not undetermined a priori constants. The reserved ports here refer to the set of reserved port nodes from step 1. .exist When invertible, construct an open-loop matrix. The sum returns the difference matrix. ,therefore and These are two ways of writing the same relation. Calculation. As an indicator of the risk frequency band distance from critical interaction; A smaller value indicates that the returned difference is close to singular, but it is not used alone to determine a stable sign.

[0047] right For frequency bands where the risk threshold is below, the passive port index is below the threshold, or a spectral peak appears in the operation monitoring, the generalized Nyquist method is used to check the number of points around the characteristic trajectory of L relative to the −1 point, or the solution is obtained. The returned complex eigenvalues ​​are obtained. The rightmost complex root is the real part of the complex root. Determine the stable sign and record the corresponding critical frequency, critical operating condition, and left and right eigenvectors.

[0048] Step 3: Determine the main equipment, candidate parameters, and responsibility.

[0049] At critical operating conditions and critical frequencies, the critical eigenvectors corresponding to the passive port parameters are used to calculate... The sensitivity of candidate control parameters to eigenvalues ​​is calculated; simultaneously, the normalized sensitivity of the rightmost complex root real part αmax to each candidate parameter is calculated using analytical derivatives or central finite difference. The root left shift sensitivity is used to determine whether parameter changes can push the critical root into the left half-plane, and the port passive index sensitivity is used to determine whether the energy characteristics of the selected device at the target frequency are improved or at least not deteriorated.

[0050] Sensitivity, allowable parameter adjustment, equipment residual current margin, modulation margin, and power margin are combined to form a responsibility level. Equipment and parameters with higher responsibility levels are included in the decision variable set in Step 4; the remaining variables remain unchanged, thus enabling sensitivity analysis to be directly used for dominant equipment location, control responsibility allocation, and optimization dimensionality reduction.

[0051] Step 4: Parameter solution of the combined constraints of port passivity and complex root damping

[0052] Instead of assuming a unique target impedance increment, it directly uses the increment of candidate control parameters. Non-negative slack variables As decision variables, with Let be the objective function, where The scaling matrix is ​​composed of the allowable adjustment amounts of the candidate parameters. This is the relaxation penalty coefficient.

[0053] Constraints include: the passive indicators of the selected devices and network ports after the update are not lower than the corresponding threshold or not lower than the indicators before the update minus the allowable amount; return the rightmost complex root real part of the difference. Not greater than ; Parameters satisfy The controller is internally stable; the converter port current, voltage, modulation ratio, additional control output, and control delay all do not exceed the equipment limits. To pre-determine the negative damping limit, parameter thresholds and operating limits are all checked according to the operating condition set Q.

[0054] Seeking Then, the difference between the reduced-order impedance matrix before and after the update is defined as the actual impedance change. .thus, It is the result of the updated constrained control parameters, not the only decision quantity that needs to be directly inferred. If the RLC local approximation is used to reduce the computational cost, this approximation is only used for generation. The initial values ​​are determined based on the port passive indices, complex root damping, and operational constraints of the original multi-port model.

[0055] Step 5: Parameter Mapping and Impedance Shaping Control of Heterogeneous Converters

[0056] Will The control loop with a predetermined structure is directly written. For grid-type converters, candidate structures include virtual impedance with high-frequency roll-off, PLL damping compensation, and voltage, current, or power feedforward links; for network-type converters, candidate structures include virtual inertia, virtual damping, droop coefficient, and frequency-selective additional damping. The optimized output is the specific parameter increment, target device number, activation time, and slope limit, and multiplicative adaptation functions with unclear relationships to parameter increments are no longer introduced.

[0057] Each new transmission link adopts a stable, real-coefficient, causal structure with an order not exceeding a preset value, and incorporates sampling, calculation, and communication delays into the controller model. After the parameters are written into the shadow model or supervisory layer, the internal stability of the controller is checked by poles or state matrix eigenvalues, and then distributed according to the slope limit and graded step size method.

[0058] Step 6: Supervise retesting, accept criteria, and rollback in case of failure.

[0059] Let the first The second verification yielded the safety parameters. and candidate increment ,according to Update, in which Indicates the set of allowed parameters projection, from The parameters are selected through backtracking, where 0 < β < 1. Candidate parameters are first verified in the shadow model or the supervision layer, and then distributed using a slope constraint.

[0060] After each update, the port admittance is re-identified within the preset observation window, the passive parameters of the selected port, the return error risk parameters, and the rightmost complex root are recalculated, and the voltage, current, power, modulation ratio, and additional control output are checked simultaneously. The system accepts updates when the target port passive indicators have not deteriorated beyond the allowable limit and all operational constraints are met; if the strict stability check fails, the port passive indicators deteriorate, the operating load exceeds the limit, communication is interrupted, or the identification confidence is insufficient, the system immediately rolls back to its previous state. and reduce .

[0061] When the above acceptance conditions are maintained continuously Updates terminate at the end of each verification cycle; maximum iteration count reached. If the conditions are still not met, online updates will be frozen and an alarm will be issued. This forms a closed-loop self-stabilizing control mechanism of "multi-port modeling - passive port and risk screening - strict stability judgment of complex roots - sensitivity responsibility positioning - constraint parameter solution - supervised retesting and rollback".

[0062] Example 1: Self-stabilizing control of approximately 124Hz oscillation in a four-section 35kV bus heterogeneous renewable energy power plant

[0063] This embodiment uses an independent copy of the existing four-bus electromagnetic transient model for verification; the original model remains unchanged. The 330kV side represents the ideal voltage boundary. The substation retains two main transformers, four 35kV busbars, collector lines, box-type transformers, and grid-connected and grid-forming equivalent converters on each busbar. The static var generator is removed from this control line to maintain consistency between the frequency domain model and the electromagnetic transient boundary. The analysis frequency band is 110–140Hz, with a frequency step of 0.25Hz.

[0064] 1. System parameters and multi-port model

[0065] The system parameters are shown in Table 1.

[0066] Table 1

[0067]

[0068] Establish the ports of each mesh-type and mesh-type converter at the steady-state operating point of the storage. Frequency coupling admittance, and device admittance formed in port order. The box-type transformer, collector lines, main transformer, and transmission line admittance are assembled into... Network node admittance matrix Eight sequence domain ports corresponding to the four 35kV busbar segments are retained as a node set. The two main transformer star nodes and the 330kV node of the power station are used as the internal node set. Eliminate the original element to obtain ; then calculate and The impedance base value is taken as... .

[0069] 2. Risk frequency bands, passive indicators at ports, and strict stability assessment

[0070] For the first Port definition The passive network was verified using the corresponding index of network admittance. The minimum eigenvalue of the Hermitian portion of the network admittance within the 110–140 Hz range is 0.002905, which is greater than 0. The return difference matrix for the baseline operating condition is shown. A critical interaction occurs near 123.84 Hz, with the real part of the rightmost complex root being +2.7479 s−1; the identification frequency in the electromagnetic transient is 124.16 Hz, and the envelope growth rate is +0.5703 s−1, which is determined to be an oscillating condition.

[0071] This embodiment also verifies an important applicability boundary: and the whole The minimum eigenvalues ​​of the Hermitian part are not guaranteed to be monotonic when crossing stability boundaries; therefore, they are only used for risk band screening and sufficiency checks, and not for determining the stable sign independently. The stable sign is determined by the returned difference complex eigenvalues ​​and verified by the electromagnetic transient envelope; the port passive index is used to limit the energy characteristics of the selected equipment at the target frequency to ensure that they do not deteriorate. Figure 2 The passive specifications of the selected port at the target frequency of 124Hz are shown.

[0072] Table 2 compares the passive performance of the selected mesh / network aggregation ports of motherboards I and III under baseline operating conditions and after constrained co-correction. The port performance improved by 0.00217 and 0.00959 respectively, indicating that the final accepted control parameters did not sacrifice the dissipation margin of the selected ports at the target frequency.

[0073] Table 2

[0074]

[0075] 3. Dominant equipment and candidate control channels

[0076] Applying the same 10% normalization reduction to the bandwidth of each of the four bus segments and the mesh-type PLL, the rightmost real parts were obtained as +0.2834, +5.9934, −1.9277, and −0.8174s−1, respectively. Although the critical mode node participation of bus IV was the highest, the PLL parameters of bus III were most effective in shifting the rightmost real part to the left, indicating that node participation cannot replace parameter sensitivity.

[0077] Table 3 lists the root left shift sensitivity, allowable variation, and responsibility share for the two selected control channels. Among them, the III-type PLL bandwidth reduction channel accounts for 60.34%, and the I-type frequency-selective active-frequency damping channel accounts for 39.66%, which are used to determine the responsibility allocation for coordinated parameter adjustment.

[0078] Table 3

[0079]

[0080] Network-type additional damping is adopted The physical positive sequence risk frequency is approximately 124Hz, corresponding to approximately 74Hz in the synchronous rotating coordinate system. Therefore, a center frequency of 73Hz is chosen. The DC gain of this bandpass circuit is 0, which does not change the steady-state droop operating point at 50Hz.

[0081] 4. Constrained parameter updates, rejections, and rollbacks

[0082] The unconstrained cooperative candidates obtained from sensitivity responsibility allocation are: the bandwidth of the III mother PLL is reduced by 0.9729%, and the additional active-frequency gain of the I mother network is approximately... The rightmost root of this candidate parameter in the frequency domain model is −0.9987s−1, but the peak value of the additional angular frequency command in the electromagnetic transient reaches approximately 950.07 rad / s and is accompanied by overmodulation risk. Therefore, the supervisory layer rejects this set of parameters and rolls it back. After the rollback, the change in the III mother PLL is retained, and the network gain is reduced to... And set a protection limit of ±5 rad / s.

[0083] Table 4 summarizes the parameters finally accepted after the supervisory layer rejected unconstrained candidates and completed the backoff. The proportional and integral coefficients of the mesh-type phase-locked loop were slightly reduced. The additional damping for the mesh-type loop was implemented with a center frequency of 73Hz, a damping ratio of 0.7, and a smooth connection method, and a protection limit of ±5 rad / s was configured.

[0084] Table 4

[0085]

[0086] 5. Verification of frequency domain complex roots and electromagnetic transient closed-loop.

[0087] Table 5 compares the frequency domain complex root and electromagnetic transient identification results for three operating conditions: reference oscillation, local correction only with mesh pattern, and constrained mesh pattern / mesh pattern coordination. The real part of the theoretical root and the electromagnetic transient growth rate of both control schemes change from positive to negative, and the identification frequency remains around 124.16 Hz, indicating that the oscillation has changed from growth to decay.

[0088] Table 5

[0089]

[0090] Figure 3 (a) The minimum singular values ​​of the return difference in the range of 110–140 Hz are given for three operating conditions. Figure 3 (b) The real part of the corresponding rightmost complex root is given. The theoretical values ​​for the three operating conditions are consistent with the electromagnetic transient stability sign, and the oscillation frequency error is less than 0.27%. The theoretical root of the constrained cooperative scheme is close to the stability boundary, but the electromagnetic transient envelope continues to decay; the peak value of the additional angular frequency output is 0.1169 rad / s, which is far below the ±5 rad / s limit, indicating that the stability is not maintained by the limiter. Figure 4 (a) shows the electromagnetic transient envelope of the risk mode. Figure 4 (b) A comparison was made between theoretical and electromagnetic transient growth rates. Figure 4 (c) gives the amplitude ratio before / after the modal envelope.

[0091] 6. Operational Constraints, Robust Conditions, and Implementation Conclusions

[0092] Figure 5 (a) The modulation peak values ​​of each grid-type and grid-type converter under the final cooperative scheme are given. The maximum values ​​of the two are 0.6854 and 0.7684, respectively, both of which are less than 1. The original grid-type modulation limiter did not operate. Figure 5(b) The steady-state active power of each 35kV bus section is given, which is approximately 124.46MW. The corresponding steady-state reactive power is approximately 1.739 Mvar, and the effective value of the 35kV bus line voltage is approximately 33.986kV. Under the non-reference operating condition where the aggregation unit of the II bus section converter is out of service and the other three sections continue to operate, the growth rate of the 124Hz component is −1.6574s−1, the maximum grid-type / grid-type modulation is 0.6942 and 0.7708 respectively, and the peak value of the additional angular frequency is 0.1401rad / s, which still meets the constraints.

[0093] Therefore, this embodiment establishes a closed-loop evidence within the limited topology, operating point, and 110–140 Hz risk band: "port passive index and return difference screening—complex root strict stability assessment—parameter sensitivity responsibility allocation—unconstrained candidate rejection—parameter backoff—electromagnetic transient and operational constraint retesting." This conclusion cannot be extrapolated to unsimulated static var generator participation, external grid impedance changes, or different output conditions; these conditions should be considered as supplementary examples in further dependent embodiments, and the values ​​in this embodiment should not be directly applied.

[0094] This invention also provides a broadband oscillation self-stabilizing control system for new energy field clusters based on passive constraints, including a processor, a memory, an impedance measurement or identification interface, an operating condition acquisition interface, and a control parameter distribution interface connected to a grid-connected converter and a network-connected converter; the memory stores instructions, which, when executed by the processor, implement the method steps in the aforementioned embodiments.

[0095] In practical applications, the aforementioned processor includes a Field-Programmable Gate Array (FPGA), and the processor can be a Central Processing Unit (CPU) or a Digital Signal Processor (DSP). It is understood that for different devices, the electronic devices used to implement the above-mentioned processor functions can also be other types, and this embodiment of the invention does not impose specific limitations.

[0096] The aforementioned memory can be volatile memory, such as random-access memory (RAM); or non-volatile memory, such as read-only memory (ROM), flash memory, hard disk drive (HDD), or solid-state drive (SSD); or a combination of the above types of memory, and provides instructions and data to the processor.

[0097] In an exemplary embodiment, the present invention also provides a computer-readable storage medium for storing a computer program.

[0098] Optionally, the computer-readable storage medium can be applied to any of the methods in the embodiments of the present invention, and the computer program causes the computer to execute the corresponding processes implemented by the processor in the various methods of the embodiments of the present invention. For the sake of brevity, these will not be described in detail here.

[0099] In the several embodiments provided by this invention, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

Claims

1. A method for broadband oscillation self-stabilization control of new energy field clusters based on passive constraints, characterized in that... Includes the following steps: Step 1, in the set of operating conditions Operating conditions in Below, using the steady-state operating points of the grid-type converter and the mesh-type converter as linearization points, the reduced-order impedance matrix of the field group multi-port is obtained. ; Step 2: Divide the frequency domain into multiple characteristic frequency bands. Define passive indicators for device ports and passive indicators of network ports Assemble the network admittance YN and the device admittance YD and construct the return difference matrix. ,by Screening for risk bands; Step 3, when the port passive indicator is below the threshold, When the value falls below the risk threshold or the return complex eigenvalue enters the right half-plane, the critical frequency, critical operating condition, and critical eigenvector are determined; the real part of the rightmost complex root and the normalized sensitivity of the port passive index to the candidate control parameters are calculated, and the dominant converter and the set of candidate control parameters are determined accordingly. Step 4, increment the candidate control parameters As the decision variable, under the constraints that the passive index of the target port is not lower than the corresponding threshold, the real part of the rightmost complex root of the return difference is not greater than the preset negative damping limit, the parameters are within the allowable range, the controller is internally stable, the voltage, current and modulation amount do not exceed the limits, and the control delay is limited, the control parameter increment that minimizes the change of the normalized parameter is solved. Step 5: Map the obtained control parameter increment to at least one of the virtual impedance, phase-locked loop damping, feedforward compensation, virtual inertia, virtual damping, or droop control loop of the dominant converter, and determine the actual impedance increment based on the difference between the reduced-order impedance matrix before and after the update. Step 6: Issue control parameters according to the update rules with projection and backtracking step size, remeasure or identify port admittance and recalculate port passive indicators, return difference eigenvalues ​​and operating constraints; roll back to the previous set of safety parameters when strict stability check fails, target port passive indicators deteriorate beyond the allowable amount or operating quantity exceeds the limit; stop updating when the complex root damping and operating constraints are met for a preset number of consecutive times.

2. The method as described in claim 1, characterized in that: The specific content of step 1 is as follows: In the set of operating conditions Operating conditions in Below, using the steady-state operating points of the grid-type converter and the mesh-type converter as linearization points, we obtain the first... Dynamic state of the converter Port voltage disturbances and port current disturbances are denoted as follows: , and ; Establish the small-signal admittance matrix of each converter port In addition to the nodal admittance matrices of the collector lines, transformers, and external power grid, internal nodes are eliminated to obtain a reduced-order impedance matrix that preserves the coupling relationship between the observation ports. .

3. The method as described in claim 1, characterized in that: In step 1, the first The port voltage disturbance and port current disturbance of the converter are represented by two-dimensional vectors in a synchronously rotating dq coordinate system, wherein the port small-signal admittance matrix is... It is obtained by linearizing the converter power stage, filter, digital control delay and control loop at the steady-state operating point; when using positive and negative order coordinates, the dq admittance matrix is ​​converted into an order domain admittance matrix containing positive and negative order coupling terms through a preset coordinate transformation matrix.

4. The method as described in claim 1, characterized in that: In step 1, the network node admittance matrix formed by the collector lines, transformers, and external power grid is arranged according to the set of reserved port nodes. and the set of internal nodes of the element to be eliminated Divided into blocks , , and ;in, This represents the self-admittance and mutual admittance submatrices between the retained port nodes. This represents the coupling submatrix of internal node voltages to the injected current at reserved ports. This represents the coupling submatrix that preserves the port voltage's relationship to the injected current at the internal nodes. Represents the self-admittance and mutual admittance submatrices between internal nodes; in When reversible, according to Obtain the network reserved port admittance, and assemble the small-signal admittances of each converter port into... and in accordance with and Obtain the reduced-order impedance matrix of the field group multi-port.

5. The method as described in claim 1, characterized in that: In step 2, the first is defined. Normalized port admittance of the converter Its Hermitian part is and according to Calculate the first The converter in the first Passive indicators of device ports in each characteristic frequency band Construct an open-loop matrix The sum returns the difference matrix. and calculate As an indicator of risk frequency band distance, the stable symbol is... The characteristic trajectory is relatively Number of points around the point or The rightmost complex root is determined.

6. The method as described in claim 1, characterized in that: In step 3, at the critical frequency and critical operating conditions, the sensitivity of the passive index of the port is calculated using the critical eigenvector, and the sensitivity of the real part of the rightmost complex root to the candidate parameters is calculated using the central finite difference method. The responsibility is formed according to the sensitivity, the allowable adjustment of the parameters, the residual current of the equipment, and the modulation margin.

7. The method as described in claim 1, characterized in that: In step 4, with The objective function is a weighted sum of slack variables, with the objective port passive index not lower than a set value after constraint update, and the real part of the rightmost complex root of the return difference not greater than a certain value. , To preset the negative damping limit and satisfy parameter boundaries, internal stability, current, voltage, modulation ratio and additional control output limits.

8. The method as described in claim 1, characterized in that: In step 5, at least one of the following parameters is adjusted for the grid-type converter: virtual impedance parameter, phase-locked loop damping compensation parameter, voltage feedforward parameter, current feedforward parameter, or power feedforward parameter; at least one of the following parameters is adjusted for the grid-type converter: virtual inertia, virtual damping, droop coefficient, or additional virtual impedance parameter; the parameter adjustment shall not change the internal stability of the controller and shall be constrained by the converter current and modulation margin.

9. The method as described in claim 1, characterized in that: In step 6, the following is adopted: Update control parameters; if the root damping does not reach the threshold, the passive performance of the port deteriorates beyond the allowable limit, the voltage, current, or modulation exceeds the limit, communication is abnormal, or the identification confidence is insufficient, cancel the update and reduce the value. .

10. A broadband oscillation self-stabilizing control system for new energy field clusters based on passive constraints, characterized in that: It includes a processor, a memory, an impedance measurement or identification interface, an operating condition acquisition interface, and a control parameter distribution interface connected to the grid-type converter and the network-type converter; the memory stores instructions, which, when executed by the processor, implement the method as described in any one of claims 1 to 9.

Images