A rapid assessment method for oscillation risk in high-proportion photovoltaic grid-connected systems

Abstract

This invention discloses a rapid assessment method for oscillation risk in high-proportion photovoltaic grid-connected systems. The method establishes a linearized small-signal model of the photovoltaic inverter; constructs a rotational transformation matrix and its inverse and derivative based on the phase angle of each photovoltaic terminal voltage; constructs a voltage impedance transformation matrix and a sensitivity matrix; loads the network node admittance matrix, divides and eliminates elements in the network node admittance matrix to obtain an equivalent admittance matrix containing only photovoltaic nodes; constructs a sensitivity matrix based on the initial sensitivity matrix and the equivalent admittance matrix; constructs a corrected AC network admittance matrix by combining the sensitivity matrix and the equivalent admittance matrix; and obtains the active axis equivalent admittance matrix after eliminating the dynamic influence of the photovoltaic system's reactive axis. The method calculates the minimum eigenvalue of the active axis equivalent admittance matrix, combines the minimum eigenvalue with the linearized small-signal model to obtain system eigenvalues, and calculates the oscillation risk index based on the real part of the system eigenvalues ​​for oscillation risk assessment. This method combines the dynamic components of the photovoltaic system with the equivalent admittance of the distribution network and directly obtains oscillation risk information by solving for the eigenvalues, achieving rapid assessment.

Description

Technical Field

[0001] This invention relates to the field of power system control technology, and in particular to a rapid assessment method for oscillation risk in high-proportion photovoltaic grid-connected systems. Background Technology

[0002] With the large-scale grid connection of renewable energy, photovoltaic power generation, as one of the main renewable energy sources, is increasingly accounting for a larger proportion of the power grid. High-proportion photovoltaic grid-connected systems have rapidly changing output and control characteristics, making grid oscillation characteristics more complex and uncertain, which affects system stability. Traditional linear stability analysis based on small disturbances can only calculate eigenvalues ​​for static linearized models, ignoring the actual dynamic characteristics of the photovoltaic inverter control loop. The analysis object is usually limited to a single inverter or low-proportion scenario, making it difficult to effectively assess the dynamic interaction effects when large-scale photovoltaic access is implemented. Although time-domain simulation methods are accurate, they require a lot of computational resources and time, making them difficult to apply in real time in actual operation. Existing methods struggle to balance real-time performance, accuracy, and applicability. Summary of the Invention

[0003] In view of this, the present invention proposes a rapid assessment method for oscillation risk of high-proportion photovoltaic grid-connected systems. Based on real-time system operation data, the method rapidly obtains system oscillation risk indicators by constructing a photovoltaic equivalent impedance network and solving for characteristic roots.

[0004] The technical solution of this invention is implemented as follows:

[0005] A rapid assessment method for oscillation risk in high-proportion photovoltaic grid-connected systems includes the following steps:

[0006] Step S1: Establish a linearized small-signal model of the photovoltaic inverter in the dq coordinate system;

[0007] Step S2: Based on the phase angle of each photovoltaic terminal voltage, construct the rotation transformation matrix from the xy coordinate system to the dq coordinate system, as well as its inverse transformation matrix, derivative of the transformation matrix, and derivative of the inverse transformation matrix;

[0008] Step S3: Construct the voltage impedance transformation matrix, and combine the derivatives of the rotational transformation matrix and the inverse transformation matrix to construct the initial sensitivity matrices of voltage and current to phase angle changes;

[0009] Step S4: Load the network node admittance matrix, divide and eliminate the network node admittance matrix to obtain the equivalent admittance matrix containing only photovoltaic nodes, construct the sensitivity matrix based on the initial sensitivity matrix and the equivalent admittance matrix, construct the corrected AC network admittance matrix by combining the sensitivity matrix and the equivalent admittance matrix, and obtain the active axis equivalent admittance matrix after eliminating the dynamic influence of the reactive axis of the photovoltaic system.

[0010] Step S5: Calculate the minimum eigenvalue of the active axis equivalent admittance matrix, combine the minimum eigenvalue with the linearized small-signal model to obtain the system characteristic roots, and calculate the oscillation risk index based on the real part of the system characteristic roots to assess the oscillation risk.

[0011] Preferably, the calculation formula for the linearized small-signal model is as follows:

[0012] ;

[0013] in , For the current loop PI controller gain, This refers to the per-unit value of the photovoltaic DC-side capacitor. , They are respectively Initial values ​​of shaft current and voltage, These are characteristic roots.

[0014] Preferably, the specific steps of step S2 are as follows:

[0015] Based on the voltage phase angle at each photovoltaic unit terminal Construct the rotation transformation matrix from the xy coordinate system to the dq coordinate system, expressed as:

[0016] ;

[0017] Construct the inverse transformation matrix from the dq coordinate system to the xy coordinate system, expressed as:

[0018] ;

[0019] Construct the derivative of the transformation matrix to calculate the partial derivative of the rotation transformation matrix with respect to the phase angle. The expression is:

[0020] ;

[0021] Construct the derivative of the inverse transformation matrix and calculate the partial derivative of the inverse transformation matrix with respect to the phase angle. The expression is:

[0022] .

[0023] Preferably, the specific steps of step S3 are as follows:

[0024] Based on the voltage measured in the xy coordinate system by the system and voltage in the dq coordinate system Constructing the voltage impedance transformation matrix The expression is:

[0025] ;

[0026] in Represents voltage in the xy coordinate system and in the dq coordinate system The real part, Represents voltage in the xy coordinate system The imaginary part;

[0027] Construct the initial sensitivity matrices of voltage and current to phase angle changes. The expressions are as follows:

[0028] ,

[0029] ,

[0030] ;

[0031] in and These are the voltage vector in the dq coordinate system and the output current vector in the xy coordinate system, respectively.

[0032] Preferably, the specific steps of step S4 are as follows:

[0033] Load the network node admittance matrix. After completing the node sorting, divide the node admittance matrix into a photovoltaic node part and other PQ node parts. The expression is:

[0034] ;

[0035] in Let PQ be the admittance matrix between nodes. and These are the admittance matrices between the PQ node and the photovoltaic node, and the admittance matrix between the photovoltaic node and the PQ node, respectively. The admittance matrix between photovoltaic nodes;

[0036] Based on nodal admittance matrix Eliminating the PQ nodes yields the equivalent admittance matrix containing only photovoltaic nodes. Its expression is:

[0037] ;

[0038] Constructing the sensitivity matrix The expression is:

[0039] ;

[0040] The corrected AC network admittance matrix is ​​constructed by combining the sensitivity matrix and the equivalent admittance matrix. :

[0041] ;

[0042] The admittance matrix of the communication network After eliminating the reactive axis effect of the photovoltaic system, the blocks are as follows:

[0043] ;

[0044] The dynamic influence of the reactive axis of the photovoltaic system is eliminated by eliminating the admittance matrix of the partitioned AC network, and the equivalent admittance matrix of the active axis of the photovoltaic system is calculated. Its expression is:

[0045] .

[0046] Preferably, the specific steps of step S5 are as follows:

[0047] according to Calculate the equivalent admittance matrix of the active axis. The smallest eigenvalue EIG;

[0048] Solving for the closed-loop characteristic roots of the system With frequency domain conversion, the calculation formula is:

[0049] ;

[0050] in For linearized small-signal models;

[0051] Calculate the oscillation risk index based on the real part of the eigenvalues:

[0052] .

[0053] Preferred options also include:

[0054] Step S6: Predict the risk diffusion path, screen the risk diffusion paths to obtain the risk diffusion path with the highest maintenance cost, and take oscillation suppression measures first based on the risk diffusion path with the highest maintenance cost.

[0055] Preferably, step S6 includes the following specific steps:

[0056] Step S61: Obtain the power grid topology diagram containing each photovoltaic node and mark the initial oscillation risk index;

[0057] Step S62: Collect photovoltaic power output data and load data, and predict the oscillation risk index at different times;

[0058] Step S63: Randomly select a photovoltaic node as the initial oscillation point. Based on the system eigenvalues ​​and eigenvectors, and combined with the power grid topology, generate the risk diffusion path of the risk caused by the initial oscillation point in the photovoltaic grid-connected system.

[0059] Step S64: Use the Grey Wolf optimization algorithm to filter all risk diffusion paths based on the highest maintenance cost, and obtain the risk diffusion path with the highest maintenance cost;

[0060] Step S65: Prioritize oscillation suppression measures for the risk diffusion path with the highest maintenance cost.

[0061] Preferably, the specific steps of step S64 are as follows: initialize the gray wolf population and the number of iterations, wherein the number of iterations is the number of risk diffusion paths, randomly select a risk diffusion path, calculate the fitness value with the highest maintenance cost, select another risk diffusion path, calculate the fitness value, compare it with the fitness value calculated in the previous iteration, retain the risk diffusion path with the larger fitness value, perform iterative calculation, obtain the risk diffusion path with the largest fitness value, and output it as the risk diffusion path with the highest maintenance cost.

[0062] Preferably, the oscillation suppression measures include adjusting the control parameters of key inverters on the risk diffusion path or installing damping devices.

[0063] Compared with the prior art, the beneficial effects of the present invention are:

[0064] This invention discloses a rapid assessment method for oscillation risk in high-proportion photovoltaic grid-connected systems. The method establishes an equivalent model of the photovoltaic grid-connected system, including models of grid nodes, photovoltaic nodes, and controllers. Then, a synchronous reference frame coordinate transformation is used to unify the variables of the photovoltaic inverter and the grid into the same coordinate system, constructing a small-signal state matrix or admittance matrix. By solving for the system's eigenvalues ​​and their damping ratios, the frequency and attenuation characteristics of each oscillation mode are obtained. Finally, each mode is quantitatively assessed according to defined oscillation risk assessment indicators, thereby rapidly determining the system's oscillation stability risk level. By combining the dynamic components of the photovoltaic system with the equivalent admittance of the distribution network and directly obtaining oscillation risk information by solving for eigenvalues, rapid assessment is achieved. Attached Figure Description

[0065] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only preferred embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0066] Figure 1 This is a flowchart of a rapid assessment method for oscillation risk in a high-proportion photovoltaic grid-connected system according to the present invention;

[0067] Figure 2This is a flowchart of another embodiment of the rapid assessment method for oscillation risk of a high-proportion photovoltaic grid-connected system according to the present invention;

[0068] Figure 3 This is a detailed step diagram of step S6 in the rapid assessment method for oscillation risk of a high-proportion photovoltaic grid-connected system according to the present invention; Detailed Implementation

[0069] To better understand the technical content of this invention, a specific embodiment is provided below, and the invention will be further described in conjunction with the accompanying drawings.

[0070] See Figure 1 This invention provides a rapid assessment method for oscillation risk in high-proportion photovoltaic grid-connected systems, comprising the following steps:

[0071] Step S1: Establish a linearized small-signal model of the photovoltaic inverter in the dq coordinate system;

[0072] Step S2: Based on the phase angle of each photovoltaic terminal voltage, construct the rotation transformation matrix from the xy coordinate system to the dq coordinate system, as well as its inverse transformation matrix, derivative of the transformation matrix, and derivative of the inverse transformation matrix;

[0073] Step S3: Construct the voltage impedance transformation matrix, and combine the derivatives of the rotational transformation matrix and the inverse transformation matrix to construct the initial sensitivity matrices of voltage and current to phase angle changes;

[0074] Step S4: Load the network node admittance matrix, divide and eliminate the network node admittance matrix to obtain the equivalent admittance matrix containing only photovoltaic nodes, construct the sensitivity matrix based on the initial sensitivity matrix and the equivalent admittance matrix, construct the corrected AC network admittance matrix by combining the sensitivity matrix and the equivalent admittance matrix, and obtain the active axis equivalent admittance matrix after eliminating the dynamic influence of the reactive axis of the photovoltaic system.

[0075] Step S5: Calculate the minimum eigenvalue of the active axis equivalent admittance matrix, combine the minimum eigenvalue with the linearized small-signal model to obtain the system characteristic roots, and calculate the oscillation risk index based on the real part of the system characteristic roots to assess the oscillation risk.

[0076] In this invention, a rapid assessment method for oscillation risk in a high-proportion photovoltaic grid-connected system first establishes several equivalent models of the photovoltaic grid-connected system, including models of grid nodes, photovoltaic nodes, and controllers. Then, a synchronous reference coordinate transformation is used to unify the variables of the photovoltaic inverter and the grid into the same coordinate system, constructing a linearized small-signal model in the dq coordinate system to achieve dynamic characteristic modeling of the photovoltaic voltage disturbance to current response. The phase angle of the terminal voltage of each photovoltaic unit is obtained, and a rotational transformation matrix and its inverse transformation matrix, the derivative of the transformation matrix, and the derivative of the inverse transformation matrix are constructed respectively. Then, a voltage impedance transformation matrix and a sensitivity matrix are constructed. The voltage impedance transformation matrix is ​​obtained by separating the real and imaginary parts of the voltage measured in the xy coordinate system and the voltage in the dq coordinate system. The sensitivity matrix needs to be combined with the voltage impedance transformation matrix and the rotational transformation matrix. The invention constructs an active-axis equivalent admittance matrix by transforming the matrix, using the derivatives of the transformation matrix and the inverse transformation matrix. After construction, the network admittance matrix of the photovoltaic grid-connected system is loaded, and several simplification steps are performed to obtain the active-axis equivalent admittance matrix. Finally, the minimum eigenvalue of the active-axis equivalent admittance matrix is ​​calculated, and the system eigenvalues ​​are obtained by combining it with a linearized small-signal model. The oscillation risk index can be calculated based on the real part of the system eigenvalues ​​to assess the oscillation risk, thereby quickly determining the oscillation stability risk level of the system. Based on real-time system data, this invention utilizes the dynamic model of the photovoltaic inverter and network equivalent analysis to construct a modified AC admittance matrix, calculate the minimum eigenvalue, and combine it with the eigenvalues ​​to calculate the oscillation risk index. This avoids complex time-domain simulation and allows for online real-time assessment of the oscillation risk of high-proportion photovoltaic grid-connected systems, with advantages such as high assessment efficiency and good accuracy.

[0077] Preferably, the calculation formula for the linearized small-signal model is as follows:

[0078] ;

[0079] in , For the current loop PI controller gain, This refers to the per-unit value of the photovoltaic DC-side capacitor. , They are respectively Initial values ​​of shaft current and voltage, These are characteristic roots.

[0080] Based on the control structure of the photovoltaic grid-connected system, the equivalent dynamic link of the active axis of the photovoltaic inverter can be obtained, and a linearized small-signal model can be constructed.

[0081] Preferably, the specific steps of step S2 are as follows:

[0082] Based on the voltage phase angle at each photovoltaic unit terminal Construct the rotation transformation matrix from the xy coordinate system to the dq coordinate system, expressed as:

[0083] ;

[0084] Construct the inverse transformation matrix from the dq coordinate system to the xy coordinate system, expressed as:

[0085] ;

[0086] Construct the derivative of the transformation matrix to calculate the partial derivative of the rotation transformation matrix with respect to the phase angle. The expression is:

[0087] ;

[0088] Construct the derivative of the inverse transformation matrix and calculate the partial derivative of the inverse transformation matrix with respect to the phase angle. The expression is:

[0089] .

[0090] Based on the voltage phase angle at each photovoltaic unit terminal We can first construct a rotation transformation matrix, and the coordinate system of the rotation transformation matrix changes from the xy coordinate system to the dq coordinate system. Therefore, the coordinate system of the inverse of the rotation transformation matrix, i.e. the inverse transformation matrix, changes from the dq coordinate system to the xy coordinate system. At the same time, in order to use it for sensitivity analysis, we construct the derivative of the transformation matrix based on the rotation transformation matrix to calculate the partial derivative of the rotation transformation matrix with respect to the phase angle. Finally, we construct the derivative of the inverse transformation matrix to calculate the partial derivative of the inverse transformation matrix with respect to the phase angle.

[0091] Preferably, the specific steps of step S3 are as follows:

[0092] Based on the voltage measured in the xy coordinate system by the system and voltage in the dq coordinate system Constructing the voltage impedance transformation matrix The expression is:

[0093] ;

[0094] in Represents voltage in the xy coordinate system and in the dq coordinate system The real part, Represents voltage in the xy coordinate system The imaginary part;

[0095] Further considering the dynamic coupling effect of the photovoltaic inverter, an initial sensitivity matrix for voltage and current to phase angle changes is constructed according to the definition. This is used to describe the effect of phase angle disturbance on the output voltage; the initial sensitivity matrix. The expressions are as follows:

[0096] ,

[0097] ,

[0098] ;

[0099] in and These are the voltage vector in the dq coordinate system and the output current vector in the xy coordinate system, respectively.

[0100] Preferably, the specific steps of step S4 are as follows:

[0101] Load the network node admittance matrix. After completing the node sorting, divide the node admittance matrix into a photovoltaic node part and other PQ node parts. The expression is:

[0102] ;

[0103] in Let PQ be the admittance matrix between nodes. and These are the admittance matrices between the PQ node and the photovoltaic node, and the admittance matrix between the photovoltaic node and the PQ node, respectively. The admittance matrix between photovoltaic nodes;

[0104] Based on nodal admittance matrix Eliminating the PQ nodes yields the equivalent admittance matrix containing only photovoltaic nodes. This can significantly reduce the system order. The expression is:

[0105] ;

[0106] Furthermore, it can be based on the initial sensitivity matrix. A complete phase angle sensitivity matrix is ​​synthesized to quantify the impact of phase angle disturbances on the overall system. The constructed sensitivity matrix... The expression is:

[0107] .

[0108] In obtaining Then, based on the above definitions, the modified AC network admittance matrix is ​​constructed. :

[0109] ;

[0110] AC network admittance matrix Taking into account both the active dynamic coupling of the photovoltaic nodes themselves and the influence of the equivalent network admittance, this is an equivalent admittance description of the photovoltaic system in the AC coordinate system.

[0111] To assess the oscillation risk, the admittance matrix of the AC network needs to be... Eliminate the reactive axis effect of the photovoltaic system, The blocks are as follows:

[0112] ;

[0113] The dynamic influence of the reactive axis of the photovoltaic system is eliminated by eliminating the admittance matrix of the partitioned AC network, and the equivalent admittance matrix of the active axis of the photovoltaic system is calculated. Its expression is:

[0114] .

[0115] Preferably, the specific steps of step S5 are as follows:

[0116] according to Calculate the equivalent admittance matrix of the active axis. The smallest eigenvalue EIG;

[0117] Solving for the closed-loop characteristic roots of the system With frequency domain conversion, the calculation formula is:

[0118] ;

[0119] in For linearized small-signal models;

[0120] Calculate the oscillation risk index based on the real part of the eigenvalues:

[0121] .

[0122] In this definition, the risk index approaches 0% when the real part of the eigenvalue approaches 0; and the risk index is 100% when the real part is -2. By setting a threshold, the damping level and oscillation risk of the system can be quickly determined. This invention realizes the direct calculation of oscillation risk through network admittance and photovoltaic dynamic model, providing an effective tool for stability monitoring of high-proportion photovoltaic grid-connected systems.

[0123] All the steps described above can be implemented using a power grid monitoring system and mathematical calculation software without the need to build an actual physical experimental platform. In practical implementation, the algorithm can be deployed in the power grid dispatch and control center to acquire and analyze real-time data from high-proportion photovoltaic grid-connected systems. This method has advantages such as low computational load and fast response speed, making it suitable for online dynamic stability monitoring. Compared with traditional small-signal analysis and time-domain simulation methods, this invention improves the evaluation speed and real-time performance. By using eigenvalues ​​for rapid solution without stepwise time-domain simulation, this method can complete oscillation risk assessment in seconds or even less, greatly improving the real-time performance of monitoring. By considering the full dynamic control behavior of photovoltaic inverters, including the impact of current loop, voltage loop, and PLL on system stability, the evaluation results are closer to actual operating conditions. It can simultaneously model and analyze grid-connected systems with multiple inverters, and is equally effective for complex systems with high proportions and multiple access points, solving the problem of the inability to scale traditional single-machine analysis. By extracting key eigenvalues ​​and conducting targeted power grid damping characteristic assessments, high-risk oscillation modes can be effectively identified, guiding subsequent control and regulation measures and enhancing the system's oscillation suppression capability.

[0124] Reference Figure 2-3 Another embodiment shown also includes:

[0125] Step S6: Predict the risk diffusion path, screen the risk diffusion paths to obtain the risk diffusion path with the highest maintenance cost, and take oscillation suppression measures first based on the risk diffusion path with the highest maintenance cost.

[0126] Another embodiment of the present invention provides specific steps after oscillation risk assessment. The main purpose is to predict possible risk diffusion paths and assess and compare the maintenance costs of each risk diffusion path to obtain the risk diffusion path with the highest maintenance cost. The risk diffusion path with the highest maintenance cost represents a greater impact on the photovoltaic grid-connected system and will require more manpower, material resources and other resources for maintenance. Therefore, it has the highest priority and oscillation suppression measures can be taken first for the risk diffusion path with the highest maintenance cost to avoid risk diffusion.

[0127] Preferably, step S6 includes the following specific steps:

[0128] Step S61: Obtain the power grid topology diagram containing each photovoltaic node and mark the initial oscillation risk index;

[0129] Step S62: Collect photovoltaic power output data and load data, and predict the oscillation risk index at different times;

[0130] Step S63: Randomly select a photovoltaic node as the initial oscillation point. Based on the system eigenvalues ​​and eigenvectors, and combined with the power grid topology, generate the risk diffusion path of the risk caused by the initial oscillation point in the photovoltaic grid-connected system.

[0131] Step S64: Use the Grey Wolf optimization algorithm to filter all risk diffusion paths based on the highest maintenance cost, and obtain the risk diffusion path with the highest maintenance cost;

[0132] Step S65: Prioritize oscillation suppression measures for the risk diffusion path with the highest maintenance cost.

[0133] The acquired power grid topology map contains several photovoltaic (PV) nodes. The calculated oscillation risk index can be used as the initial oscillation risk index for each PV node and marked accordingly. As time changes, the oscillation risk index will also change. If risk spreads, the oscillation risk index of the subsequently spreading nodes will not be the initial oscillation risk index. Therefore, this embodiment combines PV output data and load data to predict the oscillation risk index of different nodes at different times. Then, all PV nodes are used as initial oscillation points. Based on the system eigenvalues ​​and eigenvectors, and combined with the power grid topology map, the risk spread path in the PV grid-connected system caused by the initial oscillation point can be obtained. One initial oscillation point may spread to multiple paths, and multiple initial oscillation points will form a risk spread path of a certain scale. In order to achieve rapid path selection, this embodiment introduces the Grey Wolf optimization algorithm, which aims to maximize maintenance cost. It optimizes and selects all risk spread paths. After obtaining the risk spread path with the highest maintenance cost, oscillation suppression measures can be prioritized to avoid risk spread and ensure the stable operation of the PV grid-connected system.

[0134] Preferably, the specific steps of step S64 are as follows: initialize the gray wolf population and the number of iterations, wherein the number of iterations is the number of risk diffusion paths, randomly select a risk diffusion path, calculate the fitness value with the highest maintenance cost, select another risk diffusion path, calculate the fitness value, compare it with the fitness value calculated in the previous iteration, retain the risk diffusion path with the larger fitness value, perform iterative calculation, obtain the risk diffusion path with the largest fitness value, and output it as the risk diffusion path with the highest maintenance cost.

[0135] When selecting paths, the Grey Wolf Optimization Algorithm iteratively calculates the fitness of each risk diffusion path and compares them. It retains the risk diffusion path with the larger fitness value in each comparison. Finally, through multiple comparisons, it can retain the risk diffusion path with the largest fitness value, which is the risk diffusion path with the highest maintenance cost.

[0136] Preferably, the oscillation suppression measures include adjusting the control parameters of key inverters on the risk diffusion path or installing damping devices.

[0137] Adjusting the control parameters of key inverters along the risk propagation path can quickly change the inverter's output impedance characteristics, break the original impedance matching resonance condition, directly reduce the oscillation amplitude and frequency, and reduce power fluctuations caused by continuous oscillation. Adding damping devices can suppress the transmission of oscillation signals in specific frequency bands by consuming oscillation energy or adjusting the system's equivalent impedance, thus preventing the risk from spreading along the propagation path.

[0138] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A rapid assessment method for oscillation risk in a high-proportion photovoltaic grid-connected system, characterized in that, Includes the following steps: Step S1: Establish a linearized small-signal model of the photovoltaic inverter in the dq coordinate system; Step S2: Construct a rotation transformation matrix from the xy coordinate system to the dq coordinate system based on the phase angle of each photovoltaic terminal voltage. and its inverse transformation matrix Derivatives of transformation matrices Inverse transformation matrix derivative ; Step S3: Construct the voltage impedance transformation matrix By combining the derivatives of the rotational transform matrix and the inverse transform matrix, initial sensitivity matrices for voltage and current to phase angle changes are constructed. ; Step S4: Load the network node admittance matrix, divide and eliminate the network node admittance matrix to obtain the equivalent admittance matrix containing only photovoltaic nodes, construct the sensitivity matrix based on the initial sensitivity matrix and the equivalent admittance matrix, construct the corrected AC network admittance matrix by combining the sensitivity matrix and the equivalent admittance matrix, and obtain the active axis equivalent admittance matrix after eliminating the dynamic influence of the reactive axis of the photovoltaic system. Step S5: Calculate the minimum eigenvalue of the active axis equivalent admittance matrix, combine the minimum eigenvalue with the linearized small-signal model to obtain the system eigenvalues, and calculate the oscillation risk index based on the real part of the system eigenvalues ​​to assess the oscillation risk. The specific steps of step S4 are as follows: Load the network node admittance matrix. After completing the node sorting, divide the node admittance matrix into a photovoltaic node part and other PQ node parts. The expression is: ； in Let PQ be the admittance matrix between nodes. and These are the admittance matrices between the PQ node and the photovoltaic node, and the admittance matrix between the photovoltaic node and the PQ node, respectively. The admittance matrix between photovoltaic nodes; Based on nodal admittance matrix Eliminating the PQ nodes yields the equivalent admittance matrix containing only photovoltaic nodes. Its expression is: ； Constructing the sensitivity matrix The expression is: ； The corrected AC network admittance matrix is ​​constructed by combining the sensitivity matrix and the equivalent admittance matrix. : ； The admittance matrix of the communication network After eliminating the reactive axis effect of the photovoltaic system, the blocks are as follows: ； The dynamic influence of the reactive axis of the photovoltaic system is eliminated by eliminating the admittance matrix of the partitioned AC network, and the equivalent admittance matrix of the active axis of the photovoltaic system is calculated. Its expression is: 。 2. The method for rapid assessment of oscillation risk in a high-proportion photovoltaic grid-connected system according to claim 1, characterized in that, The calculation formula for the linearized small-signal model is as follows: ； in , For the current loop PI controller gain, This refers to the per-unit value of the photovoltaic DC-side capacitor. , They are respectively Initial values ​​of shaft current and voltage, These are characteristic roots.

3. The method for rapid assessment of oscillation risk in a high-proportion photovoltaic grid-connected system according to claim 1, characterized in that, The specific steps of step S2 are as follows: Based on the voltage phase angle at each photovoltaic unit terminal Construct the rotation transformation matrix from the xy coordinate system to the dq coordinate system, expressed as: ； Construct the inverse transformation matrix from the dq coordinate system to the xy coordinate system, expressed as: ； Construct the derivative of the transformation matrix to calculate the partial derivative of the rotation transformation matrix with respect to the phase angle. The expression is: ； Construct the derivative of the inverse transformation matrix and calculate the partial derivative of the inverse transformation matrix with respect to the phase angle. The expression is: 。 4. The method for rapid assessment of oscillation risk in a high-proportion photovoltaic grid-connected system according to claim 3, characterized in that, The specific steps of step S3 are as follows: Based on the voltage measured in the xy coordinate system by the system and voltage in the dq coordinate system Constructing the voltage impedance transformation matrix The expression is: ； in Represents voltage in the xy coordinate system and in the dq coordinate system The real part, Represents voltage in the xy coordinate system The imaginary part; Construct the initial sensitivity matrices of voltage and current to phase angle changes. The expressions are as follows: ， ， ； in and These are the voltage vector in the dq coordinate system and the output current vector in the xy coordinate system, respectively.

5. The method for rapid assessment of oscillation risk in a high-proportion photovoltaic grid-connected system according to claim 1, characterized in that, Also includes: Step S6: Predict the risk diffusion path, screen the risk diffusion paths to obtain the risk diffusion path with the highest maintenance cost, and take oscillation suppression measures first based on the risk diffusion path with the highest maintenance cost.

6. The method for rapid assessment of oscillation risk in a high-proportion photovoltaic grid-connected system according to claim 5, characterized in that, The specific steps of step S6 include: Step S61: Obtain the power grid topology diagram containing each photovoltaic node and mark the initial oscillation risk index; Step S62: Collect photovoltaic power output data and load data, and predict the oscillation risk index at different times; Step S63: Randomly select a photovoltaic node as the initial oscillation point. Based on the system eigenvalues ​​and eigenvectors, and combined with the power grid topology, generate the risk diffusion path of the risk caused by the initial oscillation point in the photovoltaic grid-connected system. Step S64: Use the Grey Wolf optimization algorithm to filter all risk diffusion paths based on the highest maintenance cost, and obtain the risk diffusion path with the highest maintenance cost; Step S65: Prioritize oscillation suppression measures for the risk diffusion path with the highest maintenance cost.

7. The method for rapid assessment of oscillation risk in a high-proportion photovoltaic grid-connected system according to claim 6, characterized in that, The specific steps of step S64 are as follows: Initialize the gray wolf population and the number of iterations, where the number of iterations is the number of risk diffusion paths. Randomly select a risk diffusion path and calculate the fitness value with the highest maintenance cost. Select another risk diffusion path and calculate its fitness value. Compare it with the fitness value calculated in the previous step. Keep the risk diffusion path with the larger fitness value. After iterative calculation, obtain the risk diffusion path with the largest fitness value and output it as the risk diffusion path with the highest maintenance cost.

8. The method for rapid assessment of oscillation risk in a high-proportion photovoltaic grid-connected system according to claim 6, characterized in that, The oscillation suppression measures include adjusting the control parameters of key inverters on the risk diffusion path or installing damping devices.

Images