Frequency safety assessment method and device for photovoltaic grid-connected power system and electronic equipment

Abstract

The invention discloses a photovoltaic grid-connected power system frequency safety evaluation method and device and electronic equipment, and the method comprises the steps: carrying out the grouping of all photovoltaic power generation units based on the operation state, initial operation power, and grid-connected point voltage under an expected fault of the photovoltaic power generation units, and obtaining a plurality of photovoltaic power generation units with different active responses; performing parameter equivalence calculation on the divided photovoltaic power generation units of various types, aggregating clusters by applying a capacity weighted average method, and fitting the characteristics of the active slope recovery clusters by adopting a two-section slope function; performing parameter equivalence calculation on the divided photovoltaic power generation units to obtain an active analysis formula of the photovoltaic power generation units in the low voltage ride through process; combining the active analytic expression with a traditional SFR frequency response model to obtain a photovoltaic unit frequency response model considering the low voltage ride through behavior; and solving the photovoltaic unit frequency response model to obtain frequency response. According to the method, the influence of the LVRT active characteristic on the frequency safety index is analyzed.

Description

Technical Field

[0001] This invention relates to the field of power systems, and more particularly to a method, apparatus, and electronic device for frequency security assessment of photovoltaic grid-connected power systems that takes into account low voltage ride-through characteristics. Background Technology

[0002] With the widespread grid connection of large-scale photovoltaic (PV) power generation units, my country is accelerating its low-carbon energy transition and the formation of a new power system, which has become one of the main ways to achieve its dual-carbon goals. The construction of this new power system requires ample security as a technological guarantee. PV power plants, characterized by converter grid-connected control, typically operate in maximum power point tracking (MPPT) mode. When disturbances occur, they do not respond to unbalanced power changes and are decoupled from the system frequency. Therefore, the large-scale grid connection of PV systems reduces the system's inertia level and frequency regulation capability, posing a significant challenge to the system's frequency security.

[0003] Meanwhile, when a power system accident or disturbance causes a significant voltage drop at the grid connection point of a photovoltaic (PV) power plant, the PV power plant should possess low voltage ride-through (LVRT) capability. During LVRT, the PV power plant follows the reactive current priority principle, reducing active power injection to maximize voltage support capacity. As seen throughout the LVRT process, the PV power plant participates in the frequency response of the sending-end system through active power changes when the grid connection point voltage drops significantly. How to account for the PV LVRT control behavior and establish a frequency response model for a large-scale PV grid-connected system is of significant research value for accurately describing the system's frequency security.

[0004] Power system frequency response modeling methods are mainly divided into three categories: time-domain simulation, data-driven methods, and approximate solution methods. Time-domain simulation represents the system as a system of differential-algebraic equations and uses numerical integration to obtain the frequency response. However, this method suffers from high computational cost and low solution efficiency, making it difficult to apply to large-scale photovoltaic grid-connected systems. Data-driven methods establish online evaluation models based on data obtained from frequency events, but this method is limited by the sample size of the frequency event data. Approximate solution methods retain the active power components such as the generator rotor and governor, and use capacity weighting and frequency regulation factor weighting to aggregate all power sources and reduce their order to a second-order system frequency response (SFR) model. Based on this, SFR models for high-proportion renewable energy power systems are also obtained by weighting renewable energy generation units according to the principle of capacity weighting. However, these models neglect the active power variation of photovoltaic LVRT behavior.

[0005] The key is to establish an analytical formula for the active power of LVRT that can accurately describe the actual photovoltaic power source. Therefore, in the case of low voltage ride-through of photovoltaic new energy, the study of frequency stability of AC and DC systems will become a research focus. Summary of the Invention

[0006] This invention provides a method, device, and electronic equipment for frequency security assessment of photovoltaic grid-connected power systems. The invention achieves photovoltaic cluster division based on the active power response differences of photovoltaic power generation units (LVRTs), aggregates each cluster using a capacity-weighted average method, and uses a two-segment slope function to fit the characteristics of the active power slope recovery cluster. Finally, it derives the equivalent active power analytical expression for the entire LVRT process of the photovoltaic power plant to reduce the order of the SFR model and improve the solution rate. Subsequently, it combines the traditional SFR model with the low-voltage ride-through dynamic process of the photovoltaic power plant to establish a frequency response model for a large-scale photovoltaic power plant power system that considers voltage-active power-frequency dynamics, as detailed below:

[0007] A first aspect: a method for frequency security assessment of a photovoltaic grid-connected power system, the method comprising:

[0008] Based on the operating status, initial operating power, and grid connection point voltage under anticipated faults, all photovoltaic power generation units are grouped to obtain a variety of photovoltaic power generation units with different active power responses.

[0009] The parameters of the various types of photovoltaic power generation units are calculated by equivalent calculation. The capacity-weighted average method is applied to aggregate the clusters. The active power slope recovery characteristics of the clusters are fitted by a two-segment slope function. The active power analytical expressions of various types of photovoltaic power generation units during the low voltage ride-through process are obtained.

[0010] By combining the active power analytical expression with the traditional SFR frequency response model, a photovoltaic unit frequency response model considering low voltage ride-through behavior is obtained.

[0011] The frequency response is obtained by solving the frequency response model of the photovoltaic unit.

[0012] The active power expression is:

[0013] ;

[0014] in, Pde GroupB(t) and Pde GroupC(t) are the active power analytical expressions for clusters of types A, B, and C, respectively.

[0015] Among them, the active power expression for type A clusters is:

[0016] ;

[0017] In the formula, nA1 n A2 Let n represent the number of photovoltaic power generation units in group A whose initial active power is less than or equal to Pavg_A norm and greater than Pavg_A norm, respectively. A1 + n A2 =n; teqA cturn is the transition point of the two-stage recovery function, that is, the active power is restored to n. A1 The moment when the initial active power within the group is summed; teqA cend is the end time of the recovery phase of the two-stage recovery function, i.e., the active power is restored to n. A The moment when the initial active power in the group is summed; k represents the active current recovery rate, ieqA dmax(t) is the maximum allowable d-axis current of the Class A photovoltaic unit during the low-voltage duration, and uPCC flt(t) is the output voltage I of the PCC node during the low-voltage duration. max The maximum output current limit of the inverter is ieqA*q_flt(t). The q-axis current of the Class A photovoltaic unit during the low-voltage duration is uPCC norm(t). The output voltage PPVi norm(t) during normal operation of the PCC node is PPVi. The active power of the photovoltaic node during normal operation is q_flt(t).

[0018] Among them, the active power analytical expressions for type B and type C clusters are:

[0019] ;

[0020] ;

[0021] In the formula, The superscript eq represents the equivalent unit, n represents the total number of photovoltaic units in the group, t0 is the start time of low voltage at the photovoltaic unit port, and t c t is the time when the photovoltaic port voltage recovers. cend When the active power output of the photovoltaic system recovers to its initial value, [t0,t] c This indicates a sustained low-pressure phase; u norm u flt These represent the output voltage of the photovoltaic power generation unit during the fault-free operation phase and the low-voltage continuous phase, respectively; id out_norm The d-axis current reference value represents the fault-free operation phase; id maxThe maximum allowable d-axis current of the inverter during the low-voltage duration phase is given by uPCC flt(t), the output voltage of the PCC node during the low-voltage duration, ieqB dmax(t) is the maximum allowable d-axis current of the Class B photovoltaic unit during the low-voltage duration, uPCC norm(t) is the output voltage of the PCC node during normal operation, ieqB dout_norm(t) is the reference value of the d-axis current of the Class B photovoltaic unit during the fault-free operation phase, PPVi norm(t) is the active power of the photovoltaic node during normal operation, and ieqC dout_norm(t) is the reference value of the d-axis current of the Class C photovoltaic unit during the fault-free operation phase.

[0022] The photovoltaic cell frequency response model considering low-voltage ride-through behavior is as follows:

[0023] ;

[0024] In the formula, This represents the transition point in the two-stage fitting process. This marks the end of the photovoltaic low-voltage ride-through process. Let represent the frequency time-domain variation of the photovoltaic grid-connected system in the i-th time period.

[0025] The method further includes: assessing the safety based on the frequency response and in conjunction with frequency security requirements, and repairing photovoltaic power generation units with safety issues.

[0026] Secondly, a frequency security assessment device for a photovoltaic grid-connected power system considering low-voltage ride-through characteristics, the device comprising:

[0027] The grouping module is used to group all photovoltaic power generation units based on their operating status, initial operating power, and grid connection point voltage under anticipated faults, thereby obtaining a variety of photovoltaic power generation units with different active power responses.

[0028] The first acquisition module is used to perform parameter equalization calculations on the various types of photovoltaic power generation units, apply the capacity weighted average method to aggregate each cluster, and use a two-segment slope function to fit the active power slope recovery characteristics of the cluster; and obtain the active power analytical expressions of various types of photovoltaic power generation units during the low voltage ride-through process.

[0029] The second acquisition module is used to combine the active power analytical expression with the traditional SFR frequency response model to obtain a photovoltaic unit frequency response model that considers low voltage ride-through behavior.

[0030] The maintenance module is used to solve the frequency response model of the photovoltaic unit to obtain the frequency response.

[0031] The device further includes: assessing the safety based on the frequency response and in conjunction with frequency security requirements, and repairing photovoltaic power generation units with safety issues.

[0032] Thirdly, an electronic device for frequency security assessment of a photovoltaic grid-connected power system, the electronic device comprising: a processor and a memory, the memory storing program instructions, the processor calling the program instructions stored in the memory to cause the device to perform the method described in any of the first aspects.

[0033] Fourth aspect, a computer-readable storage medium storing a computer program, the computer program including program instructions that, when executed by a processor, cause the processor to perform the method described in any one of the first aspects.

[0034] The beneficial effects of the technical solution provided by this invention are:

[0035] 1. This invention constructs a photovoltaic unit LVRT active response category partition boundary to accurately and efficiently divide photovoltaic clusters, uses the capacity weighting method to calculate the equivalent unit structural parameters, and uses a two-segment slope fitting method to approximately characterize the nonlinear recovery characteristics of the low-passage active slope recovery cluster.

[0036] 2. The present invention provides an analytical formula for the equivalent active power output of a photovoltaic power plant at the LVRT. By combining the traditional SFR model with the active power dynamics of the LVRT of a photovoltaic power plant, a system frequency response model considering the change in active power output of the photovoltaic power plant under the scenario of voltage drop at the grid connection point is established, which provides rich theoretical value for practical engineering research.

[0037] 3. The evaluation method proposed in this invention considers factors such as the degree of fault drop and the proportion of photovoltaic power output in the lighting scenario, and analyzes the impact of LVRT active power characteristics on frequency security indicators.

[0038] 4. This invention can provide power system operation and dispatch personnel with a more complete and sufficient theoretical identification system; at the same time, it can provide technical support for the dynamic stability monitoring and control of the power system and carry out maintenance on photovoltaic power generation units with safety problems. Attached Figure Description

[0039] Figure 1 A flowchart of a frequency security control method for a photovoltaic grid-connected power system that considers low-voltage ride-through characteristics;

[0040] Figure 2 A zoning map of the active power response categories of photovoltaic power generation units (LVRTs);

[0041] Figure 3 A detailed comparison of the active power curves of Class A clusters with equivalent LVRT;

[0042] Figure 4 For the frequency response model of the power system;

[0043] Figure 5 This is a schematic diagram of the active power deficit of the LVRT in a photovoltaic power plant after the equivalent calculation.

[0044] Figure 6 This is a topology diagram of a photovoltaic power station;

[0045] (a) is the equivalent circuit diagram; (b) is the topology diagram.

[0046] Figure 7 For U PCC Comparison of dynamic response curves when the drop depth is 0.3 pu;

[0047] The top figure shows the system frequency response; the bottom figure shows the photovoltaic active power response.

[0048] Figure 8 For U PCC Comparison of dynamic response curves after a drop depth of 0.6 PU;

[0049] The top figure shows the system frequency response; the bottom figure shows the photovoltaic active power response.

[0050] Figure 9 This is a statistical chart showing the error calculation results under different fault severity levels. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be described in further detail below.

[0052] Example 1

[0053] A frequency security assessment method for photovoltaic grid-connected power systems considering low-voltage ride-through characteristics, see [link to relevant documentation]. Figure 1 The method includes the following steps:

[0054] 101: Based on the operating status, initial operating power, and grid connection voltage under anticipated faults, all photovoltaic power generation units are grouped into three categories.

[0055] The three categories are: Category A curves, where active power is reduced during the sustained low-voltage phase and recovers to the pre-fault level with a fixed slope during the power recovery phase; Category B curves, where active power is reduced during the sustained low-voltage phase and recovers to the pre-fault level instantaneously during the power recovery phase; and Category C curves, where active power output remains unchanged compared to the pre-fault level during the sustained low-voltage phase.

[0056] 102: Perform parameter equivalent calculations on the three types of photovoltaic power generation units to obtain the active power analytical formulas for the three types of photovoltaic power generation units during the low voltage ride-through process;

[0057] Since the active power deficit of the Type A curve has a significant impact on the frequency, this embodiment of the invention uses a two-segment ramp function to fit the active power analytical expression for Type A photovoltaic power generation units.

[0058] 103: By combining the active power analytical expression with the traditional SFR frequency response model, a photovoltaic unit frequency response model considering low voltage ride-through behavior is obtained;

[0059] 104: Solve the frequency response model of the photovoltaic unit considering low voltage ride-through behavior to obtain the frequency response, and evaluate the safety in combination with frequency safety requirements.

[0060] In summary, the embodiments of the present invention, through steps 101-104 above, perform fault severity adaptability verification and initial power adaptability verification, verifying that the frequency response curves of the SFR model proposed in the embodiments of the present invention have a high degree of overlap with the detailed model. This indicates that the proposed method can accurately identify and classify photovoltaic power generation units with different LVRT active power categories, and the constructed LVRT equivalent active power analytical formula can accurately reflect the active power dynamic process of the photovoltaic power station. In addition, the proposed SFR model can simulate the full-process frequency dynamic response of the detailed model under different fault severity levels. The embodiments of the present invention improve the accuracy of active power modeling of photovoltaic systems during low-voltage ripple, and realize power system frequency stability security assessment based on the SFR model considering photovoltaic low-voltage ripple, thereby improving the safety of the power system and reducing the occurrence of accident rates.

[0061] Example 2

[0062] The scheme in Example 1 will be further described below with specific calculation formulas and examples:

[0063] 201: Group all photovoltaic units based on their operating status, initial operating power, and grid connection voltage under anticipated faults;

[0064] A photovoltaic (PV) power generation unit is a typical power generation device connected to the grid via a full-power converter. Its active power response at the output is the result of the combined effect of its port voltage and active current. Based on this, the dynamic active power response P for the entire LVRT process is written. PV (t) is:

[0065] (1)

[0066] In the formula, t0 is the starting time of the low voltage at the photovoltaic unit port, t c t is the time when the photovoltaic port voltage recovers. cendWhen the active power output of the photovoltaic system recovers to its initial value, let [~,t0] and [t] be denoted as t0, ... c [,~] represents the fault-free operation phase, [t0,t c This indicates a sustained low-pressure phase; u norm u flt These represent the output voltage of the photovoltaic power generation unit during the fault-free operation phase and the low-voltage continuous phase, respectively; id out_norm id out_flt These represent the d-axis current reference values ​​calculated from the voltage outer loop during the fault-free operation phase and the low-voltage continuous operation phase, respectively; id max This represents the maximum allowable d-axis current output of the inverter during the low-voltage continuous phase; k represents the active current recovery rate.

[0067] Observing equation (1), it can be seen that the active power characteristics of the LVRT of the photovoltaic unit are mainly related to the fault duration voltage and the initial active power, and may exhibit the following characteristics: Figure 2 The three distinct active power dynamic characteristics shown (hereinafter referred to as type A, type B, and type C) are compared by examining the initial active power P of the photovoltaic power generation unit. norm The maximum active power output Pmax flt and t during the low-voltage continuous period c Output power P(t) at any time c The relationship between these three factors allows for a direct determination of the LVRT active power category of the photovoltaic unit. The determination principles and standards are as follows:

[0068] 1) Category A. For example... Figure 2 As shown by the curve in region A on the left, active power is reduced during the sustained low-voltage phase, and during the power recovery phase, active power recovers to the pre-fault level with a fixed slope. If P norm >P(t c If the LVRT active power category of the photovoltaic power generation unit is A, then the active power category of the photovoltaic power generation unit is A.

[0069] 2) Category B. For example... Figure 2 As shown by the curve in the middle B-type region, active power is reduced during the sustained low-voltage phase, and the active power instantaneously recovers to the pre-fault level during the power recovery phase. If P(t) c )≥P norm If Pmax flt, then the LVRT active power category of this photovoltaic power generation unit is Class B;

[0070] 3) Category C. For example... Figure 2 As shown by the curve in region C on the right, the active power output remains unchanged compared to the pre-fault level during the sustained low-voltage phase. If Pmax flt ≥ P norm If so, the active power category of the LVRT of the photovoltaic power generation unit is Class C.

[0071] Generally speaking, the low-pressure phase is extremely short, and it can be approximated that the duration of this phase is within a short period. flt Unchanged, at the same time, tc Voltage recovery completed at time (u) PV (t c )=1p.u.), then Pmax flt, P(t) c They can be calculated separately as follows:

[0072] (2)

[0073] In the formula, I max This represents the maximum output current limit of the inverter, and K represents the dynamic reactive current regulation coefficient of the inverter.

[0074] Observing equation (2), we can see that Pmax flt and P(t) c All of these are only about u flt The function, i.e., P norm u flt It is a key factor affecting the active power category of LVRT photovoltaic power generation units. In the embodiments of this invention, I is taken as... max Given 1.1pu and K=1.5, we can obtain Pmaxflt and P(t). c Regarding u flt The functional relationship between the photovoltaic power generation unit LVRT active response category and zoning is as follows: Figure 2 As shown.

[0075] Given the initial active power of the photovoltaic power generation unit and the fault duration voltage, it is possible to combine Figure 2 The active LVRT response category of the photovoltaic unit is clearly known, and it is then classified into a photovoltaic cluster with the same response category for further equivalent calculation.

[0076] 202: Based on the parameter equivalent calculation of three types of photovoltaic units, the active power analytical formula of photovoltaic power station during low voltage ride-through is obtained;

[0077] Typically, the photovoltaic units within a large-scale photovoltaic power plant are all identical. In this case, the main equivalent parameters can be calculated as follows:

[0078] (3)

[0079] In the formula, N s N p These represent the number of series and parallel connections in the photovoltaic array, respectively; I rr_PVi S represents the operating solar irradiance of the i-th photovoltaic power generation unit; PV R represents the capacity of the photovoltaic power generation unit. f L f C dc These are the filter resistor, filter inductor, and DC link capacitor of the photovoltaic power generation unit, respectively; R T L T ST The resistance, reactance, and capacity of the box-type transformer are represented by the subscript "eq". The subscript "eq" represents the equivalent photovoltaic power generation unit, and n represents the total number of photovoltaic power generation units in the same group.

[0080] The active response of a photovoltaic power plant's LVRT is the superposition of the active responses of a massive number of photovoltaic power generation units, that is:

[0081] (4)

[0082] In the formula, Pde PVPP(t) represents the active power analysis formula for the entire LVRT process of a photovoltaic power station; Pde GroupA(t), Pde GroupB(t), and Pde GroupC(t) are the detailed active power analysis formulas for the entire LVRT process of photovoltaic units in groups A, B, and C, respectively.

[0083] In practical engineering, photovoltaic power generation units within a photovoltaic power station are mostly connected in parallel and have short electrical distances between them. It is reasonably assumed that the port voltage of each photovoltaic unit is equal to the voltage at the point of common coupling (PCC) of the power station. PCC Under this assumption, the active power analytical expressions Pde GroupB(t) and Pde GroupC(t) for clusters B and C can be expressed as follows:

[0084] (5)

[0085] (6)

[0086] In the formula, The superscript eq represents an equivalent unit, and n represents the total number of photovoltaic units in the group.

[0087] The duration of the active power recovery phase for Class A photovoltaic power generation units varies significantly due to the influence of the initial active power distribution effect. This causes the detailed active power analytical expression for Group A to exhibit nonlinear characteristics during the power recovery phase, which existing traditional equivalent methods cannot reflect, resulting in substantial errors. A comparison of the two is shown in the figure below. Figure 3 As shown.

[0088] Figure 3 In this context, Pde GroupB(t), Ptra_eq GroupB(t), and Peq GroupB(t) represent the detailed active power analysis formula for Class A clusters, the traditional equivalent active power analysis formula, and the active power analysis formula proposed in this embodiment of the invention, respectively. Figure 3It can be seen that the more pronounced the initial active power distribution effect of the photovoltaic power generation units in Group A, the greater the error of the traditional equivalent method. To balance calculation accuracy and efficiency, we consider using the average initial active power of all photovoltaic power generation units in Group A, Pavg_A norm, as the boundary, and construct a two-segment recovery function to approximate the active power recovery behavior of the detailed analytical expression. A schematic diagram of the recovery behavior is shown below. Figure 3 The analytical expression for the active power response (Peq GroupA(t)) of the entire LVRT process of the equivalent unit of a Class A cluster can be expressed as:

[0089] (7)

[0090] In the formula, n A1 n A2 Let n represent the number of photovoltaic power generation units in group A whose initial active power is less than or equal to Pavg_A norm and greater than Pavg_A norm, respectively. A1 + n A2 =n; teqA cturn is the transition point of the two-stage recovery function, that is, the active power is restored to n. A1 The moment when the initial active power within the group is summed; teqA cend is the end time of the recovery phase of the two-stage recovery function, i.e., the active power is restored to n. A The moment when the initial active power of the group is summed.

[0091] In summary, the dynamic analysis of active power throughout the LVRT process of a photovoltaic power plant can be equivalently represented by equations (5), (6), and (7):

[0092] (8)

[0093] Compared to Equation (4), in Equation (8), the equivalent units B and C are constant known values ​​in each stage of LVRT, and the equivalent unit A is also approximated by a two-stage linearization method. Therefore, the difficulty of solving the problem is significantly reduced, which effectively simplifies the active power dynamic solution process of photovoltaic power plants in the SFR model and improves the solution calculation efficiency.

[0094] 203: By combining the active power analytical expression of the photovoltaic unit during the low voltage ride-through process with the traditional SFR frequency response model, a photovoltaic unit frequency response model considering low voltage ride-through behavior is obtained.

[0095] Assume the set of nodes in the power system is Ω = {i ∈ N | i ≤ n} = Ω G ∪Ω PV ∪Ω D , where Ω G For the set of synchronous generator nodes, Ω PV A collection of photovoltaic power plants; Ω LLet n be the set of load nodes; n be the number of system nodes; and N be the set of natural numbers. The frequency response of a power system is mainly determined by the inertia level and primary frequency regulation capability. The active-frequency equivalent elements of all power sources are projected onto the assumed center of inertia (COI), and its rotor motion equations are established as follows:

[0096] (9)

[0097] Aggregating the primary frequency regulation participation of all power sources in the system according to the droop coefficient and capacity ratio yields the system's frequency response model, as shown below. Figure 4 As shown.

[0098] Figure 4 In the middle, T RH F is the equivalent reheat time constant. HP R is the equivalent volume factor, D is the equivalent droop factor, and K is the system damping factor.

[0099] According to the SFR model, the frequency time-domain analytical expression can be obtained by using the inverse Laplace transform:

[0100] (10)

[0101] The equivalent active power analytical expression for the entire LVRT process of a photovoltaic power plant is written out. The traditional SFR model is combined with the low voltage ride-through dynamic process of the photovoltaic power plant to establish a frequency response model of a power system with a large-scale photovoltaic power plant that takes into account voltage-active power-frequency dynamics.

[0102] Equation (8) further represents the change in active power of the photovoltaic power station ΔP throughout the entire LVRT process. PVPP (t) is:

[0103] (11)

[0104] Simultaneously, combining the active power variation characteristics of the three equivalent units in equation (8), the time-series characteristics of the active power deficit of the power station are described as follows: Figure 5 Show.

[0105] Considering the extremely short recovery time of the grid voltage, t is ignored. c0 -t c During the transition process, the active power surge during the low-voltage sustained phase can be decomposed into an event occurring at time t0 (with an amplitude of ΔP). PVPP (t0)) and t c Time (amplitude is -ΔP) PVPP (t c The superposition of two step responses; the power recovery process can be decomposed into occurring at t cThe input to the SFR model is the superposition of the four ramp responses at time teqActurn, teqAcend, and teqActurn. Therefore, the input can be considered as the superposition of six components (including the active component caused by the fault):

[0106] (12)

[0107] In the formula,

[0108] (13)

[0109] Since the SFR model is a linear model (satisfying the superposition theorem), the frequency deviation of the SFR model output can be solved by superimposing the frequency deviation components corresponding to each input component:

[0110] (14)

[0111] in,

[0112] (15)

[0113] In the formula, ε(t) is the unit step input; h(t) is the unit step response of the SFR model; and c(t) is the unit ramp response of the SFR model.

[0114] Combining equations (14) and (15), we get Δf(t) as:

[0115] (16)

[0116] 204: The main quantitative indicators of system frequency security include the maximum rate of frequency change, the maximum frequency offset, and the frequency quasi-steady-state deviation. The analysis primarily focuses on the impact of voltage-active power coupling dynamics on these three frequency security indicators.

[0117] (1) Maximum rate of change of frequency

[0118] Considering that the generator speed regulation system is not activated during the initial stage of active power disturbance in a photovoltaic power station (at time t=t0), the system frequency deviation is 0, and the frequency change rate takes the maximum value R. max ,Right now:

[0119] (17)

[0120] As can be seen from equation (17), apart from the system's inertia, the maximum frequency change rate of the system is related to the active power consumed at the fault point and the change in photovoltaic active power. If the power consumed at the fault point is greater and the change in photovoltaic output is greater, the maximum frequency change rate will be greater.

[0121] (2) Maximum frequency offset

[0122] By taking the piecewise derivative of the frequency deviation function Δf(t), we obtain the derivative function Δf′(t). Then, we search for feasible solutions t1, t2, ..., tt' that satisfy Δf′(t) = 0 in each time interval of Δf′(t). n Then the maximum frequency offset Δf max for:

[0123] (18)

[0124] (3) Frequency quasi-steady-state deviation

[0125] Considering that the active power of the system reaches equilibrium again, and the system frequency also reaches the quasi-steady-state value, the quasi-steady-state frequency deviation Δf of the system can be calculated. qss for:

[0126] (19)

[0127] As can be seen from equation (19), if the dynamic change of active power throughout the entire LVRT process of the photovoltaic power station is 0, then the quasi-steady-state deviation of the system frequency is 0. This is because the photovoltaic power station has active power recovery capability, and after the recovery phase ends, the active power output of the power station remains consistent with that before the fault.

[0128] In summary, the active power dynamics of the entire LVRT process of a photovoltaic power station have no effect on the quasi-steady-state frequency deviation of the system, but they do affect the maximum frequency change rate and the maximum frequency offset of the system. The degree of influence is closely related to the power consumed at the fault point when the low voltage fault occurs, the instantaneous active power impact of the photovoltaic power station, and the active power recovery rate of the power station after the fault is cleared.

[0129] Example 3

[0130] This section uses a 5-node test system containing a photovoltaic power plant as an example to verify the accuracy and applicability of the proposed frequency response modeling method for large-scale photovoltaic grid-connected systems considering low-voltage ride-through characteristics. A simulation system is built in MATLAB / Simulink software as follows: Figure 6 The 5-node test system containing a photovoltaic power station is shown.

[0131] Figure 6 The example system shown includes one conventional synchronous generator unit and one photovoltaic power station. The rated capacity of the conventional generator unit is 220 MVA, and the rated capacity of the photovoltaic power station is 100 MVA. The photovoltaic power station consists of 32 identical photovoltaic power generation units, and their parameters are shown in [reference needed]. .

[0132] Table 1 Simulation parameters of photovoltaic power generation unit

[0133]

[0134] A three-phase short-circuit ground fault occurred at node 2 at simulation time t=35s, and was cleared after 0.15s. The accuracy and applicability of the proposed modeling method were verified under different voltage drop / fault severity scenarios and different initial active power distributions. The proposed model was compared and verified with a detailed time-domain simulation model of a photovoltaic power plant, a traditional SFR model, and a single-cell SFR model. To verify the applicability of the proposed modeling method to fault severity, voltage drops of 0.1pu, 0.2pu, 0.3pu, 0.4pu, 0.5pu, 0.6pu, 0.7pu, 0.8pu, and 0.9pu were set at the PCC point of the photovoltaic power plant, and the solutions for the above four types of models were calculated respectively. Due to space limitations, this section uses PCC voltage drops of 0.3pu and 0.6pu as examples to analyze the performance of the proposed modeling method. The photovoltaic power generation unit clustering results under different fault severity are shown in Table 2.

[0135] Table 2U PCC Clustering results for drop depths of 0.3 pu and 0.6 pu

[0136]

[0137] As shown in Table 2, the classification results of the LVRT active response of photovoltaic power generation units differ under different voltage drop depths. No Class C units appear when the voltage drop is deep, which verifies the effectiveness of the proposed classification boundary.

[0138] Based on the above classification results, the frequency response of the four types of models and the active power output curves of photovoltaic power plants are compared as follows: Figure 7 and Figure 8 As shown.

[0139] Combination Figure 7 and Figure 8 It is evident that, under different voltage dip depths, the traditional SFR model, which does not consider the active power dynamics of the LVRT in photovoltaic power plants, performs the worst in simulating the frequency response curve of the detailed time-domain simulation model. The single-unit SFR model represents the photovoltaic power plant equivalent to a single photovoltaic unit, but because this method cannot account for the differences in active power categories of the LVRT within the photovoltaic power plant, it still cannot accurately simulate the frequency response curve of the detailed model. The SFR model proposed in this embodiment has a high degree of overlap with the frequency response curve of the detailed model, indicating that the proposed method can accurately identify and classify photovoltaic power generation units with different LVRT active power categories, and that the constructed LVRT equivalent active power analytical expression can accurately reflect the active power dynamic process of the photovoltaic power plant. Furthermore, the proposed SFR model can simulate the entire process of the detailed model's frequency response dynamics under different fault severity levels. Simultaneously, it can be seen that the system's maximum frequency change rate and maximum frequency deviation decrease with increasing voltage dip depth, mainly because...

[0140] When the voltage drops by 0.6 pu, the active power at the fault point is 19 MW, which is much smaller than the 114 MW active power when the voltage drops by 0.3 pu, while the instantaneous changes in photovoltaic LVRT are 73 MW and 23 MW, respectively. Overall, the system exhibits the characteristic of "the greater the voltage drop, the smaller the active power deficit" during faults, which contributes to frequency stability.

[0141] Calculate the maximum rate of change error of the frequency response of the three models compared with the time-domain simulation model under two sets of fault severity.

[0142] The calculation results of the maximum frequency offset error are shown in Table 3. As can be seen from Table 3, the deviation of the modeling method proposed in this embodiment is the smallest under different fault severity levels, proving that the proposed modeling method has high accuracy and strong applicability to different fault severity levels.

[0143] Table 3. Frequency index error statistics when UPCC drop depth is 0.3pu and 0.6pu

[0144]

[0145] Furthermore, using the frequency response curve of the detailed model as a benchmark, an average error index e is defined to quantify the errors of the three models. The error evaluation index e is calculated as follows:

[0146] (20)

[0147] In the formula, N represents the total number of sampling points, i represents the number of sampling points, and f TDS f SFR These represent the frequency values ​​of the time-domain simulation model and the SFR model, respectively.

[0148] The sampling interval for error calculation is from the period before the voltage drop to the period after the detailed model frequency stabilizes at 50Hz. Error calculation results for different fault severity levels are shown below. Figure 9 .

[0149] Depend on Figure 9It is evident that the traditional SFR model, which does not consider the dynamic LVRT of the photovoltaic power station, exhibits the largest error under different voltage drop depths, reaching a maximum of 0.1455 Hz. Compared to the traditional model, the error of the traditional single-unit SFR model is significantly reduced. However, when the voltage drop depth at the PCC point of the photovoltaic power station exceeds 0.7 pu, the error of this model still increases significantly, reaching a maximum of 0.0333 Hz. This is mainly because when the voltage drop is deep, the LVRT characteristics of the photovoltaic power station gradually become dominant in influencing the model accuracy. The proposed SFR model controls the average error within 0.014 Hz under different PCC point fault voltages, fully demonstrating that the proposed modeling method has strong applicability to fault severity. The constructed evaluation index can be used to accurately assess the frequency security of the photovoltaic grid-connected power system, reducing the accident rate of the power system.

[0150] Example 4

[0151] A frequency security assessment device for photovoltaic grid-connected power systems considering low-voltage ride-through characteristics, the device comprising:

[0152] The grouping module is used to group all photovoltaic power generation units based on their operating status, initial operating power, and grid connection point voltage under anticipated faults, thereby obtaining a variety of photovoltaic power generation units with different active power responses.

[0153] The first acquisition module is used to perform parameter equalization calculations on the various types of photovoltaic power generation units, apply the capacity weighted average method to aggregate each cluster, and use a two-segment slope function to fit the active power slope recovery characteristics of the cluster; and obtain the active power analytical expressions of various types of photovoltaic power generation units during the low voltage ride-through process.

[0154] The second acquisition module is used to combine the active power analytical expression with the traditional SFR frequency response model to obtain a photovoltaic unit frequency response model that considers low voltage ride-through behavior.

[0155] The maintenance module is used to solve the frequency response model of the photovoltaic unit to obtain the frequency response.

[0156] The device also includes: assessing the safety based on the frequency response and in conjunction with frequency security requirements, and repairing photovoltaic power generation units with safety issues.

[0157] The active power expression is:

[0158] ;

[0159] in, Pde GroupB(t) and Pde GroupC(t) are the active power analytical expressions for clusters of types A, B, and C, respectively.

[0160] In summary, the embodiments of the present invention improve the accuracy of active power modeling of photovoltaic systems during low-level shaving, realize power system frequency stability security assessment based on the SFR model considering photovoltaic low-level shaving, improve power system security, and reduce the occurrence of accident rates.

[0161] Example 5

[0162] An electronic device for frequency security assessment of a photovoltaic grid-connected power system considering low-voltage ride-through characteristics, the electronic device comprising: a processor and a memory, wherein the memory stores program instructions, and the processor invokes the program instructions stored in the memory to cause the device to perform the following method steps in Embodiment 1:

[0163] Based on the operating status, initial operating power, and grid connection point voltage under anticipated faults, all photovoltaic power generation units are grouped to obtain a variety of photovoltaic power generation units with different active power responses.

[0164] The parameters of the various types of photovoltaic power generation units are calculated by equivalent calculation. The capacity-weighted average method is applied to aggregate the clusters. The active power slope recovery characteristics of the clusters are fitted by a two-segment slope function. The active power analytical expressions of various types of photovoltaic power generation units during the low voltage ride-through process are obtained.

[0165] By combining the active power analytical expression with the traditional SFR frequency response model, a photovoltaic unit frequency response model considering low voltage ride-through behavior is obtained.

[0166] The frequency response is obtained by solving the frequency response model of the photovoltaic unit.

[0167] The active power expression is:

[0168] ;

[0169] in, Pde GroupB(t) and Pde GroupC(t) are the active power analytical expressions for clusters of types A, B, and C, respectively.

[0170] Among them, the active power expression for type A clusters is:

[0171] ;

[0172] In the formula, n A1 n A2 Let n represent the number of photovoltaic power generation units in group A whose initial active power is less than or equal to Pavg_A norm and greater than Pavg_A norm, respectively. A1 + n A2 =n; teqA cturn is the transition point of the two-stage recovery function, that is, the active power is restored to n. A1The moment when the initial active power within the group is summed; teqA cend is the end time of the recovery phase of the two-stage recovery function, i.e., the active power is restored to n. A The moment when the initial active power in the group is summed; k represents the active current recovery rate, ieqA dmax(t) is the maximum allowable d-axis current of the Class A photovoltaic unit during the low-voltage duration, and uPCC flt(t) is the output voltage I of the PCC node during the low-voltage duration. max The maximum output current limit of the inverter is ieqA*q_flt(t). The q-axis current of the Class A photovoltaic unit during the low-voltage duration is uPCC norm(t). The output voltage PPVi norm(t) during normal operation of the PCC node is PPVi. The active power of the photovoltaic node during normal operation is q_flt(t).

[0173] Among them, the active power analytical expressions for B and C type clusters are:

[0174] ;

[0175] ;

[0176] In the formula, The superscript eq represents the equivalent unit, n represents the total number of photovoltaic units in the group, t0 is the start time of low voltage at the photovoltaic unit port, and t c t is the time when the photovoltaic port voltage recovers. cend When the active power output of the photovoltaic system recovers to its initial value, [t0,t] c This indicates a sustained low-pressure phase; u norm u flt These represent the output voltage of the photovoltaic power generation unit during the fault-free operation phase and the low-voltage continuous phase, respectively; id out_norm The d-axis current reference value represents the fault-free operation phase; id max The maximum allowable d-axis current of the inverter during the low-voltage duration phase is given by uPCC flt(t), the output voltage of the PCC node during the low-voltage duration, ieqB dmax(t) is the maximum allowable d-axis current of the Class B photovoltaic unit during the low-voltage duration, uPCC norm(t) is the output voltage of the PCC node during normal operation, ieqB dout_norm(t) is the reference value of the d-axis current of the Class B photovoltaic unit during the fault-free operation phase, PPVi norm(t) is the active power of the photovoltaic node during normal operation, and ieqC dout_norm(t) is the reference value of the d-axis current of the Class C photovoltaic unit during the fault-free operation phase.

[0177] The frequency response model for photovoltaic cells considering low-voltage ride-through behavior is as follows:

[0178] ;

[0179] In the formula, This represents the transition point in the two-stage fitting process. This marks the end of the photovoltaic low-voltage ride-through process. Let represent the frequency time-domain variation of the photovoltaic grid-connected system in the i-th time period.

[0180] This also includes: assessing safety based on frequency response and frequency safety requirements, and repairing photovoltaic power generation units with safety issues.

[0181] In summary, the embodiments of the present invention improve the accuracy of active power modeling of photovoltaic systems during low-level shaving, realize power system frequency stability security assessment based on the SFR model considering photovoltaic low-level shaving, improve power system security, and reduce the occurrence of accident rates.

[0182] It should be noted that the device descriptions in the above embodiments correspond to the method descriptions in the embodiments, and the embodiments of the present invention will not be repeated here.

[0183] The execution entities of the aforementioned processor and memory can be devices with computing functions such as computers, microcontrollers, and single-chip microcomputers. In specific implementations, the embodiments of the present invention do not limit the execution entities and can select them according to the needs of actual applications.

[0184] Data signals are transmitted between the memory and the processor via a bus, which will not be elaborated upon in this embodiment of the invention.

[0185] Based on the same inventive concept, embodiments of the present invention also provide a computer-readable storage medium, the storage medium including a stored program, which, when the program is running, controls the device where the storage medium is located to execute the method steps in the above embodiments.

[0186] The computer-readable storage medium includes, but is not limited to, flash memory, hard disk, solid-state drive, etc.

[0187] It should be noted that the description of the readable storage medium in the above embodiments corresponds to the description of the method in the embodiments, and the embodiments of the present invention will not be repeated here.

[0188] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of the present invention is generated.

[0189] A computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be stored in or transmitted through a computer-readable storage medium. A computer-readable storage medium can be any available medium accessible to a computer or a data storage device such as a server or data center that integrates one or more available media. The available medium can be magnetic or semiconductor, etc.

[0190] Unless otherwise specified, the model numbers of the various devices in this embodiment of the invention are not limited, and any device that can perform the above functions is acceptable.

[0191] Those skilled in the art will understand that the accompanying drawings are merely schematic diagrams of a preferred embodiment, and the sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0192] 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 method for frequency security assessment of a photovoltaic grid-connected power system, characterized in that, The method includes: Based on the operating status, initial operating power, and grid connection point voltage under anticipated faults, all photovoltaic power generation units are grouped to obtain a variety of photovoltaic power generation units with different active power responses. The parameters of the various types of photovoltaic power generation units are calculated by equalization. The capacity-weighted average method is applied to aggregate the clusters. The characteristics of the active slope recovery cluster are obtained by fitting the active slope of the two-segment slope function. The active analytical expressions of various types of photovoltaic power generation units in the low voltage ride-through process are obtained. By combining the active power analytical expression with the traditional SFR frequency response model, a photovoltaic unit frequency response model considering low voltage ride-through behavior is obtained. The frequency response is obtained by solving the frequency response model of the photovoltaic unit.

2. The frequency security assessment method for a photovoltaic grid-connected power system according to claim 1, characterized in that, The active power expression is: ； in, Pde GroupB(t) and Pde GroupC(t) are the active power analytical expressions for clusters of types A, B, and C, respectively.

3. The method for frequency security assessment of a photovoltaic grid-connected power system according to claim 2, characterized in that, The active power expression for a type A cluster among various types is: ； In the formula, n A1 n A2 Let n represent the number of photovoltaic power generation units in group A whose initial active power is less than or equal to Pavg_A norm and greater than Pavg_A norm, respectively. A1 + n A2 =n; teqA cturn is the transition point of the two-stage recovery function, that is, the active power is restored to n. A1 The moment when the initial active power within the group is summed; teqA cend is the end time of the recovery phase of the two-stage recovery function, i.e., the active power is restored to n. A The moment when the initial active power in the group is summed; k represents the active current recovery rate, ieqA dmax(t) is the maximum allowable d-axis current of the Class A photovoltaic unit during the low-voltage duration, and uPCC flt(t) is the output voltage I of the PCC node during the low-voltage duration. max The maximum output current limit of the inverter is ieqA* q_flt(t). The q-axis current of the Class A photovoltaic unit during the low-voltage duration is uPCC norm(t). The output voltage of the PCC node during normal operation is PPVi norm(t). The active power of the photovoltaic node during normal operation is PPVi norm(t).

4. The frequency security assessment method for a photovoltaic grid-connected power system according to claim 2, characterized in that, The active power expression for B and C type clusters among various types is as follows: ； ； In the formula, The superscript eq represents the equivalent unit, n represents the total number of photovoltaic units in the group, t0 is the start time of low voltage at the photovoltaic unit port, and t c t is the time when the photovoltaic port voltage recovers. cend When the active power output of the photovoltaic system recovers to its initial value, [t0,t] c This is the low-pressure sustained phase; u norm u flt These represent the output voltage of the photovoltaic power generation unit during the fault-free operation phase and the low-voltage continuous phase, respectively; id out_norm The d-axis current reference value represents the fault-free operation phase; id max The maximum allowable d-axis current of the inverter during the low-voltage duration phase is given by uPCC flt(t), the output voltage of the PCC node during the low-voltage duration, ieqB dmax(t) is the maximum allowable d-axis current of the Class B photovoltaic unit during the low-voltage duration, uPCC norm(t) is the output voltage of the PCC node during normal operation, ieqB dout_norm(t) is the reference value of the d-axis current of the Class B photovoltaic unit during the fault-free operation phase, PPVi norm(t) is the active power of the photovoltaic node during normal operation, and ieqC dout_norm(t) is the reference value of the d-axis current of the Class C photovoltaic unit during the fault-free operation phase.

5. The frequency security assessment method for a photovoltaic grid-connected power system according to claim 1, characterized in that, The photovoltaic cell frequency response model considering low-voltage ride-through behavior is as follows: ； In the formula, This represents the transition point in the two-stage fitting process. This marks the end of the photovoltaic low-voltage ride-through process. Let represent the frequency time-domain variation of the photovoltaic grid-connected system in the i-th time period.

6. The method for frequency security assessment of a photovoltaic grid-connected power system according to claim 1, characterized in that, The method further includes: assessing the safety based on the frequency response and in conjunction with frequency security requirements, and repairing photovoltaic power generation units with safety issues.

7. A frequency security assessment device for a photovoltaic grid-connected power system considering low-voltage ride-through characteristics, characterized in that, The device includes: The grouping module is used to group all photovoltaic power generation units based on their operating status, initial operating power, and grid connection point voltage under anticipated faults, thereby obtaining a variety of photovoltaic power generation units with different active power responses. The first acquisition module is used to perform parameter equalization calculations on the various types of photovoltaic power generation units, apply the capacity weighted average method to aggregate each cluster, and use a two-segment slope function to fit the active power slope recovery characteristics of the cluster; and obtain the active power analytical expressions of various types of photovoltaic power generation units during the low voltage ride-through process. The second acquisition module is used to combine the active power analytical expression with the traditional SFR frequency response model to obtain a photovoltaic unit frequency response model that considers low voltage ride-through behavior. The maintenance module is used to solve the frequency response model of the photovoltaic unit to obtain the frequency response.

8. The frequency security assessment device for a photovoltaic grid-connected power system considering low-voltage ride-through characteristics according to claim 7, characterized in that, The device further includes: assessing safety based on the frequency response and in conjunction with frequency security requirements, and repairing photovoltaic power generation units with safety issues.

9. An electronic device for frequency security assessment of a photovoltaic grid-connected power system, characterized in that, The electronic device includes a processor and a memory, the memory storing program instructions, the processor calling the program instructions stored in the memory to cause the device to perform the method according to any one of claims 1-6.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, the computer program including program instructions that, when executed by a processor, cause the processor to perform the method described in any one of claims 1-6.

Images