Photovoltaic grid-connected system control method and system based on low voltage ride through scene

Abstract

The invention discloses a photovoltaic grid-connected system control method and system based on a low voltage ride through scene, and relates to the technical field of photovoltaic grid-connected systems. The fast system is suitable for analyzing dynamic characteristics of the system, and the slow system is suitable for analyzing normal steady state control of the system. The repeated low-pass phenomenon is determined based on the dynamic coupling of the first slow system reduced-order model and the second slow system reduced-order model, so that the dynamic coupling of the first slow system reduced-order model and the second slow system reduced-order model is realized, and the recognition accuracy of the repeated low-pass phenomenon is ensured. An effective control measure of the photovoltaic grid-connected system is determined on the basis of comparison of a short-term voltage dynamic state after active power output is reduced and a short-term voltage dynamic state after reactive compensation, and the effective control measure is reactive compensation based on grid-connected point triggering, so that the accuracy of the effective control measure of the photovoltaic grid-connected system is improved; and an effective measure of reactive compensation based on grid-connected point triggering is introduced.

Description

Technical Field

[0001] This invention relates to the technical field of photovoltaic grid-connected systems, and more particularly to a control method and system for a photovoltaic grid-connected system based on a low-voltage ride-through scenario. Background Technology

[0002] With the development of technology, photovoltaic grid-connected systems are gradually being applied in daily life and exist in areas with a high proportion of renewable energy. In these areas, when the active power output of wind turbines increases, the turbines repeatedly enter and exit low-voltage ride-through (LVRT) conditions, causing severe voltage fluctuations at the grid connection point. This phenomenon endangers the safe and stable operation of the power system and severely restricts the output of renewable energy units. Some existing studies have simulated repeated LVRT curves in actual renewable energy sending-end power grids and proposed a system active-voltage curve (PV curve) analysis method based on the LVRT characteristics of wind turbines. Based on this, they have given optimized control strategies to suppress repeated LVRTs. However, they have not discussed the impact of reactive power on this phenomenon, and therefore cannot guarantee the accuracy of effective control measures for photovoltaic grid-connected systems. Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings of the prior art. This invention provides a control method and system for a photovoltaic grid-connected system based on a low voltage ride-through scenario.

[0004] This invention provides a control method for a photovoltaic grid-connected system based on a low-voltage ride-through scenario, comprising:

[0005] The electromechanical transient control model of the photovoltaic grid-connected system is collected, and the normal operation period, low voltage ride-through period, and low voltage ride-through recovery period of the photovoltaic grid-connected system are determined based on the division of the electromechanical transient control model.

[0006] During the normal operation of the photovoltaic grid-connected system, the electromechanical transient control model of the photovoltaic grid-connected system is decomposed into a dual time scale to output a fast system suitable for analyzing the dynamic characteristics of the system and a slow system suitable for analyzing the normal steady-state control of the system.

[0007] In the slow system, the first slow system reduced-order model during normal operation and the second slow system reduced-order model during low voltage ride-through are collected. Based on the dynamic coupling of the first and second slow system reduced-order models, the repeated low voltage ride-through phenomenon is determined, and the grid connection point of the photovoltaic grid-connected system is marked.

[0008] Collect voltage dynamics of the photovoltaic grid-connected system in the short term after a reduction in active power output and in the short term after reactive power compensation;

[0009] Effective control measures for photovoltaic grid-connected systems are determined by comparing the short-term voltage dynamics after a reduction in active power output with the short-term voltage dynamics after reactive power compensation. The effective control measure is reactive power compensation triggered by the grid connection point.

[0010] This invention provides a control system for a photovoltaic grid-connected system based on a low-voltage ride-through scenario. The control system is applied to the aforementioned control method for a photovoltaic grid-connected system based on a low-voltage ride-through scenario. The control system includes:

[0011] The partitioning module is used to collect the electromechanical transient control model of the photovoltaic grid-connected system, and to determine the normal operation period, low voltage ride-through period and low voltage ride-through recovery period of the photovoltaic grid-connected system based on the partitioning of the electromechanical transient control model.

[0012] The decomposition module is used to perform dual-time-scale decomposition of the electromechanical transient control model of the photovoltaic grid-connected system during normal operation, so as to output a fast system suitable for analyzing the dynamic characteristics of the system and a slow system suitable for analyzing the normal steady-state control of the system.

[0013] The grid connection point module is used to collect the first slow system reduced-order model during normal operation and the second slow system reduced-order model during low voltage ride-through in a slow system. Based on the dynamic coupling of the first and second slow system reduced-order models, the repeated low voltage ride-through phenomenon is determined, and the grid connection point of the photovoltaic grid-connected system is marked.

[0014] The voltage dynamics module is used to collect the voltage dynamics of the photovoltaic grid-connected system in the short term after the active power output decreases and in the short term after reactive power compensation.

[0015] The effective control measures module is used to determine the effective control measures for the photovoltaic grid-connected system based on the comparison of the voltage dynamics in the short term after the reduction of active power output and the voltage dynamics in the short term after reactive power compensation. The effective control measures are reactive power compensation triggered by the grid connection point.

[0016] Compared with the prior art, the beneficial effects of the present invention are:

[0017] In this embodiment of the invention, the electromechanical transient control model of the photovoltaic grid-connected system is acquired using the method described herein. Based on the division of this electromechanical transient control model, the normal operation period, low voltage ride-through period, and low voltage ride-through recovery period of the photovoltaic grid-connected system are determined. During the normal operation period of the photovoltaic grid-connected system, the electromechanical transient control model of the photovoltaic grid-connected system is decomposed into a dual time scale to output a fast system suitable for analyzing the dynamic characteristics of the system and a slow system suitable for analyzing the normal steady-state control of the system. In the slow system, a reduced-order model of the first slow system during normal operation and a reduced-order model of the second slow system during low voltage ride-through are acquired. Based on the dynamic coupling of the reduced-order models of the first and second slow systems, the repeated low voltage ride-through phenomenon is determined, and the grid connection point of the photovoltaic grid-connected system is marked. The introduction of a slow system suitable for analyzing the normal steady-state control of the system realizes the dynamic coupling of the reduced-order models of the first and second slow systems, ensuring the accuracy of identifying the repeated low voltage ride-through phenomenon and improving the accuracy of identifying the grid connection point of the photovoltaic grid-connected system.

[0018] Therefore, the voltage dynamics of the photovoltaic grid-connected system in the short term after the reduction of active power output and in the short term after reactive power compensation are collected. Based on the comparison of the voltage dynamics in the short term after the reduction of active power output and in the short term after reactive power compensation, effective control measures for the photovoltaic grid-connected system are determined. The effective control measure is reactive power compensation triggered by the grid connection point. This realizes the comparison of voltage dynamics in the short term after the reduction of active power output and in the short term after reactive power compensation, improves the accuracy of the effective control measures for the photovoltaic grid-connected system, and introduces an effective measure of reactive power compensation triggered by the grid connection point. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating the control method of a photovoltaic grid-connected system based on a low-voltage ride-through scenario in an embodiment of the present invention.

[0020] Figure 2 This is a schematic diagram of the structural composition of the control system of a photovoltaic grid-connected system based on a low-voltage ride-through scenario in an embodiment of the present invention;

[0021] Figure 3 The voltage waveform diagram at the grid connection point of the repeated low-voltage photovoltaic unit provided in the IEEE 3-machine 9-node example is shown in the embodiment of the present invention.

[0022] Figure 4 The diagram showing the change in operating status of an IEEE 3-machine 9-node repeating low-voltage photovoltaic unit is provided for an embodiment of the present invention.

[0023] Figure 5 The voltage waveform diagram at the grid connection point of the photovoltaic unit in the case 1354 pegase example provided in the embodiments of the present invention is shown.

[0024] Figure 6The case 1354 pegase example provided in this embodiment of the invention is a graph showing the change in the operating status of a photovoltaic unit operating under repeated low-voltage conditions. Detailed Implementation

[0025] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0026] Please see Figures 1 to 6 A control method for a photovoltaic grid-connected system based on a low-voltage ride-through scenario is provided, applicable to the control scenario of a photovoltaic grid-connected system based on a low-voltage ride-through scenario. The control method for a photovoltaic grid-connected system based on a low-voltage ride-through scenario includes:

[0027] Step S11: Collect the electromechanical transient control model of the photovoltaic grid-connected system, and determine the normal operation period, low voltage ride-through period and low voltage ride-through recovery period of the photovoltaic grid-connected system based on the division of the electromechanical transient control model;

[0028] Step S12: During the normal operation of the photovoltaic grid-connected system, perform dual time-scale decomposition on the electromechanical transient control model of the photovoltaic grid-connected system to output a fast system suitable for analyzing the dynamic characteristics of the system and a slow system suitable for analyzing the normal steady-state control of the system.

[0029] Step S13: In the slow system, collect the first slow system reduced-order model during normal operation and the second slow system reduced-order model during low voltage ride-through. Based on the dynamic coupling of the first slow system reduced-order model and the second slow system reduced-order model, determine the repeated low voltage ride-through phenomenon and mark the grid connection point of the photovoltaic grid-connected system.

[0030] Step S14: Collect the voltage dynamics of the photovoltaic grid-connected system in the short term after the reduction of active power output and in the short term after reactive power compensation;

[0031] Step S15: Based on the comparison of voltage dynamics in the short term after the reduction of active power output and voltage dynamics in the short term after reactive power compensation, determine the effective control measures for the photovoltaic grid-connected system. The effective control measures are reactive power compensation triggered by the grid connection point.

[0032] In step S11, the electromechanical transient control model of the photovoltaic grid-connected system is acquired, and the normal operation period, low voltage ride-through period, and low voltage ride-through recovery period of the photovoltaic grid-connected system are determined based on the division of the electromechanical transient control model.

[0033] In the specific implementation of this invention, the specific steps are as follows:

[0034] S111: Real-time monitoring of the photovoltaic grid-connected system, collection of multiple working data of the photovoltaic grid-connected system, and determination of the electromechanical transient control model of the photovoltaic grid-connected system based on the multiple working data of the photovoltaic grid-connected system and the control framework logic corresponding to the photovoltaic grid-connected system.

[0035] S112: Based on the detection of the electromechanical transient control model of the photovoltaic grid-connected system, multiple working paths are determined. In each working path, the path nodes of each working path are marked, and the corresponding working data is matched for each path node.

[0036] S113: Determine the period type corresponding to the working path based on the location of each path node and the working data corresponding to each path node. The period type includes normal operation period, low voltage ride-through period and low voltage ride-through recovery period.

[0037] In the embodiments of this application, the operating status of the photovoltaic grid-connected system is determined by several key parameters. The data monitored in real time includes: grid-side electrical quantities, inverter output, DC-side parameters, environmental parameters, and control commands; grid-side electrical quantities determine whether a voltage drop has occurred in the grid; inverter output ensures stable inverter operation; DC-side parameters prevent DC overvoltage or undervoltage; environmental parameters affect photovoltaic power generation efficiency; and control commands determine the inverter output strategy.

[0038] Collects multiple operating data from the photovoltaic grid-connected system: high-speed sampling (above 10kHz): rapidly changing electrical quantities such as voltage and current, used to capture transient processes (such as voltage dips); low-speed sampling (1Hz~100Hz): slowly changing parameters such as temperature and irradiance, used for long-term operation optimization; event-triggered sampling: when a voltage dip is detected (such as Upcc<0.9pu), the sampling rate is automatically increased to 20kHz to ensure accurate recording of transient processes.

[0039] To construct an electromechanical transient control model for a photovoltaic grid-connected system, multiple operating data points and corresponding control framework logic of the photovoltaic grid-connected system are introduced. These operating data points include: electrical quantities (U, I, P, Q) reflecting the system's dynamic response; equipment status (inverter temperature, DC voltage) influencing the control strategy; and control commands (Pref, Qref) determining the system's regulation mode. A normal operation model (steady-state), a low-voltage ride-through model (transient), and a recovery phase model are introduced.

[0040] Normal operation model (steady state): The inverter operates in MPPT mode and outputs maximum power; the grid voltage is stable and the reactive power Q≈0;

[0041] Low voltage ride-through model (transient): When Upcc < 0.9pu, the inverter reduces active power output (P↓) and injects reactive power (Q↑) to support the grid voltage; the DC bus voltage may fluctuate due to power imbalance.

[0042] Recovery phase model: When the grid voltage rises to above 0.9 pu, the inverter gradually resumes active power output; avoids secondary voltage drop caused by sudden power changes.

[0043] Optionally, a photovoltaic power station encounters a grid short-circuit fault, and the voltage drops to 0.5 pu. The system detects the decrease in Upcc and immediately records data such as inverter current and DC voltage. Based on historical data, a low-voltage ride-through model is matched to predict possible dynamic responses. Control decision: The inverter reduces its active power output to 30% while injecting 80% of its reactive current capacity to support the grid voltage. Recovery phase: After the fault is cleared, the voltage rises back to 0.95 pu, and the inverter gradually returns to MPPT mode.

[0044] Based on the electromechanical transient control model, the critical operating paths of the photovoltaic grid-connected system are divided into different operating stages (normal / LVRT / recovery), and key nodes are marked on each path to ensure that the operating data of each node can be accurately monitored and matched. Different paths are divided according to the voltage drop depth (e.g., 0.9 pu, 0.5 pu). The inverter adopts different control strategies in different stages (normal / LVRT / recovery), forming independent paths. At this time, when the photovoltaic power station detects that the grid voltage has dropped to 0.7 pu, the system automatically switches to the LVRT control path. The key nodes of this path include: PCC point (monitoring whether the voltage continues to drop); DC bus (preventing voltage surge due to power imbalance); and inverter PWM modulation (adjusting the output current and injecting reactive power to support the grid).

[0045] Each node needs to be matched with corresponding real-time data for subsequent analysis and control: PCC node → voltage, current, frequency; DC bus node → DC voltage, photovoltaic current; PWM modulation node → modulation wave, carrier frequency. Optionally, during LVRT, if the system detects that the DC bus voltage (Vdc) rises to 1100V (normally 800V) due to active power limitation, this data is associated with the DC bus node of the "DC side energy path"; triggering the braking resistor (Chopper) to absorb excess energy and prevent equipment overvoltage.

[0046] Based on the characteristics of node data, the system divides the operating status into three categories: normal operation, low voltage ride-through, and recovery. The criteria for normal operation is that the grid voltage is stable and there is no disturbance. The criteria for low voltage ride-through is that the grid voltage drops and grid connection needs to be maintained. The criteria for recovery is that the grid voltage returns to the normal range, but power surges must be avoided.

[0047] Optionally, a photovoltaic power station was struck by lightning, causing a voltage drop in the power grid.

[0048] Phase 1: Normal operation → LVRT entry:

[0049] Initial state: PCC voltage: 1.0 pu; Inverter output: 100% active power, 0% reactive power;

[0050] Fault occurred: A lightning strike caused a short circuit in the power grid, and the PCC voltage dropped to 0.5 pu within 20 ms;

[0051] At this time, PCC point: Upcc = 0.5pu, dU / dt = -0.3pu / s; DC bus: Vdc increases from 800V to 950V (due to excess power); Inverter: automatically switches to LVRT mode, outputting 60% reactive current; System judgment: Entering LVRT period;

[0052] Phase 2: LVRT duration:

[0053] Control actions: The inverter reduces active power to 30% and increases reactive power to 80% capacity; the DC-side Chopper starts, limiting Vdc < 1000V; the PCC voltage is stabilized at 0.5~0.6pu; the reactive current continuously supports the grid voltage;

[0054] Phase 3: LVRT → Recovery

[0055] Fault Clearance: The grid voltage recovers to 0.92 pu within 0.5 seconds; PCC point: Upcc = 0.92 pu, dU / dt = +0.2 pu / s; Inverter: Detects voltage recovery and begins to linearly increase active power (recovering 10% per minute); System determination: Entering recovery period;

[0056] Phase 4: Full Recovery

[0057] When the active power recovers to over 90% and Upcc is stable at >0.95pu, it returns to normal operation.

[0058] Cross-validation of multi-node data avoids misjudgment (such as mistaking transient disturbances for LVRT); ensures that the control strategy (reducing active power / increasing reactive power) strictly matches the grid status; gradual power recovery during the recovery phase prevents secondary impacts on the grid; optionally, in a 100MW photovoltaic power station, the LVRT trigger accuracy rate is >99.5% in actual tests; the reactive power support response time during the fault is <20ms; and there is no overshoot during the recovery process, fully complying with the requirements of GB / T 19964-2012.

[0059] In some embodiments of this application, a preset association path matching table is collected, as shown in Table 1:

[0060] Table 1: Association Path Matching Table

[0061]

[0062] Suppose that the voltage at the PCC point of a power station suddenly drops to 0.6 pu; based on the analysis of the matching table, it is known that: 1. PCC voltage = 0.6 pu → meets the LVRT condition; 2. Reactive current ratio increases from 10% to 65% → meets the LVRT condition; 3. DC voltage = 920V → meets the LVRT condition; 4. dU / dt = -0.2 pu / s → meets the LVRT condition; Result: Entering the LVRT period; Control action: Active power drops to 40%; 5. Inject 80% capacity of reactive current.

[0063] In step S12, during the normal operation of the photovoltaic grid-connected system, the electromechanical transient control model of the photovoltaic grid-connected system is decomposed into a dual time scale to output a fast system suitable for analyzing the dynamic characteristics of the system and a slow system suitable for analyzing the normal steady-state control of the system.

[0064] In the specific implementation of this invention, the specific steps are as follows:

[0065] S121: Real-time monitoring of the normal operation of the photovoltaic grid-connected system, and judgment on the electromechanical transient control model of the photovoltaic grid-connected system based on the Tikhonov fundamental theorem, in order to define whether it conforms to the range of dual time-scale decomposition;

[0066] S122: If the range of dual time-scale decomposition is met, the dual time-scale decomposition of the electromechanical transient control model of the photovoltaic grid-connected system is triggered, and the corresponding fast system and slow system are output. At this time, the fast system is suitable for analyzing the dynamic characteristics of the system; the slow system is suitable for analyzing the normal steady-state control of the system.

[0067] In the embodiments of this application, during the normal operation of the photovoltaic grid-connected system, key parameters are monitored in real time to determine whether the fast and slow dynamic separation conditions are met, providing a basis for subsequent dual time-scale decomposition (S122); confirming whether the system has obvious time-scale separation characteristics, so that the model can be safely decomposed.

[0068] Monitor two types of data: fast variables and slow variables. For fast variables, the instantaneous value of the inverter output current has a rate of change > 1000 A / s; normal range: 50Hz fundamental frequency ± 5% harmonics. For slow variables, the average DC bus voltage has a rate of change < 10V / s; normal range: 800V ± 5%. Verify the Tikhonov condition using both fast and slow variables. At this point, the following two conditions must be met simultaneously: Condition 1 (Time Scale Separation): The rate of change of the fast variable is at least 10 times that of the slow variable; for example: inverter switching frequency (10kHz) vs. MPPT regulation (1Hz) → meets a 10000-fold difference; Condition 2 (Weak Coupling): The impact of fast variable fluctuations on slow variables is < 5%; for example: current harmonics causing DC voltage fluctuations < 20V (< 2.5% of rated value) → meets the condition.

[0069] Optional, the log of a certain inverter control system: [Time] 2024-03-20 14:30:15; [Monitoring parameters] Fast variable 1 (inverter current): rate of change = 1520A / s; Fast variable 2 (PCC voltage): dU / dt = 0.05pu / s; Slow variable 1 (DC voltage): rate of change = 8V / s; Slow variable 2 (MPPT command): adjustment interval = 4.2s;

[0070] [Tikhonov verification]

[0071] 1. Time scale ratio = 1520 / 8 = 190 > 10√

[0072] 2. Current ripple causing DC fluctuation = 12V (1.5%) < 5% √

[0073] [Decision Output]

[0074] "Allow dual timescale decomposition" → triggers step S122.

[0075] Furthermore, if the range of dual-time-scale decomposition is met, the dual-time-scale decomposition of the electromechanical transient control model of the photovoltaic grid-connected system is triggered, and the corresponding fast system and slow system are output. At this time, the fast system is suitable for analyzing the dynamic characteristics of the system; the slow system is suitable for analyzing the normal steady-state control of the system. The slow system is introduced to be suitable for analyzing the normal steady-state control of the system.

[0076] At this point, when the system meets the dual-time-scale decomposition condition (S121 verification passed), the complete electromechanical transient model is decomposed into: a fast system: handling rapid dynamics at the μs to ms level (such as current surges and voltage oscillations); and a slow system: handling slow regulation at the s level (such as MPPT tracking and temperature management). In the fast system, the extracted equation is the inverter switching state equation; the corresponding physical processes are IGBT switching, PWM modulation, and current harmonics. In the slow system, the extracted equation is the DC bus voltage regulation equation; the corresponding physical processes are photovoltaic array IV curve scanning and heat dissipation control.

[0077] Optionally, in the scenario of introducing inverter sudden overcurrent protection, the fast system response (μs level) is: if the current is detected to be >110% of the rated value, the PWM duty cycle is immediately adjusted to limit the current; the action time is <100μs; the slow system coordination is: the overcurrent event is recorded, but the MPPT target remains unchanged; in the scenario of introducing MPPT adjustment when the light intensity changes gradually, the slow system response (second level) is: if the light intensity is detected to decrease by 10%, the IV curve is rescanned; the DC voltage command vdc* is updated (takes 3 seconds); the fast system coordination is: the new voltage command is automatically adapted to maintain a smooth current transition.

[0078] Specifically, a certain monitoring system outputs:

[0079] [Time] 2024-03-21 09:15:30

[0080] [Decomposition Status] ACTIVE

[0081] -Fast System:

[0082] • Current tracking error = 0.8A (<1%)

[0083] • Switching frequency = 5.01kHz

[0084] -Slow systems:

[0085] MPPT efficiency = 98.7%

[0086] DC voltage = 805V (±1%)

[0087] [Control Commands]

[0088] - Fast systems: Increase dead time to 2.2μs (harmonic suppression)

[0089] - Slow system: Increase cooling fan speed to 70%.

[0090] In step S13, in the slow system, the first slow system reduced-order model during normal operation and the second slow system reduced-order model during low voltage ride-through are collected. The repeated low voltage ride-through phenomenon is determined based on the dynamic coupling of the first slow system reduced-order model and the second slow system reduced-order model, and the grid connection point of the photovoltaic grid-connected system is marked.

[0091] In the specific implementation of this invention, the specific steps are as follows:

[0092] S131: Perform real-time monitoring of the slow system and mark the range of the slow system at different times to determine multiple first system data during normal operation and multiple second system data during low voltage ride-through.

[0093] S132: During normal operation, a reduced-order model for the first slow system is determined based on multiple first system data and the expression corresponding to the slow system; during low-voltage ride-through, a reduced-order model for the second slow system is determined based on multiple second system data and the expression corresponding to the slow system.

[0094] S133: Collect the first slow system reduced-order model and the second slow system reduced-order model, dynamically couple the first slow system reduced-order model and the second slow system reduced-order model, determine the repetition range based on the dynamic coupling of the first slow system reduced-order model and the second slow system reduced-order model, and determine the repetitive low-pass phenomenon based on the analysis of the repetition range, so as to mark the grid connection point of the photovoltaic grid-connected system.

[0095] In the embodiments of this application, during the operation of the photovoltaic grid-connected system, the slow system (the part responsible for second-level steady-state control) is monitored in real time, and the data is classified and labeled according to the grid status as: first system data (during normal operation); second system data (during low voltage ride-through); providing a structured data foundation for the subsequent establishment of a reduced-order model (S132) and repeated low voltage ride-through analysis (S133).

[0096] Monitoring key parameters of slow systems is crucial, as these parameters clearly reflect the system's status. Table 2 shows a matching table for key parameters of slow systems.

[0097] Table 2: Matching Table of Key Parameters for Slow Systems

[0098]

[0099] First system data (normal): Conditions: Upcc continuously ≥ 0.9 pu and reactive power ratio < 20%; Storage content: MPPT voltage / current, ambient temperature, irradiance; Second system data (LVRT): Conditions: Upcc < 0.9 pu and reactive power ratio ≥ 50%; Storage content: reactive power output, DC voltage fluctuation value, protection action record.

[0100] Furthermore, during normal operation, the first slow system reduction model is determined based on multiple first system data and the expression corresponding to the slow system; during low voltage ride-through, the second slow system reduction model is determined based on multiple second system data and the expression corresponding to the slow system, taking into account the overall consideration of multiple second system data and the expression corresponding to the slow system, thus ensuring the accuracy of the second slow system reduction model.

[0101] At this point, based on the data classified by S131, two dedicated order reduction models are established: the first slow system order reduction model (normal period): optimizes power generation efficiency; the second slow system order reduction model (LVRT period): ensures grid support capacity.

[0102] For the reduced-order model of the first slow system (during normal period), key parameters are extracted from the data of the first system: photovoltaic array output, environmental parameters, and inverter efficiency. The photovoltaic array output is used to establish the MPPT characteristic curve; the environmental parameters are used to correct for the effects of temperature / illuminance; and the inverter efficiency is used to evaluate energy conversion losses.

[0103] Typical expression: P = η·G·[1-0.0045(T-25)]·A; where η is the inverter efficiency (taking the historical average of 98.2%) and G is the real-time irradiance (W / m²). 2 T: Photovoltaic panel temperature (°C); A: Array area (m²) 2 Practical application: Irradiance G = 900 W / m 2 Temperature T = 45℃; Array area A = 2000m² 2 P = 0.982 × 900 × [1 - 0.0045(45 - 25)] × 2000 = 1543 kW (the error between this and the actual measured value of 1548 kW is < 0.3%).

[0104] Results: [First Slowest System Model] updated on 2024-03-26;

[0105] ├─Applicable Scenarios: U pcc ≥0.9pu;

[0106] ├─Core Function: MPPT Optimization;

[0107] └─Accuracy Verification: Historical data matching accuracy is 99.1%.

[0108] For the reduced-order model of the second slow system (during LVRT), key parameters are extracted from the second system data: reactive power support, DC voltage fluctuation, and protection action records. Reactive power support is used to calculate voltage support requirements; DC voltage fluctuation is used to assess the degree of energy surplus; and protection action records are used to optimize braking resistor control.

[0109] Typical expression: Q = min(Qmax, 0.8·(1-U / 0.9)·Srated); Qmax: maximum no-power capacity of the inverter (e.g., 400kVar); Srated: rated apparent power (e.g., 500kVA); Q = min(400, 0.8×(1-0.5 / 0.9)×500) = min(400, 178) = 178kVar;

[0110] Structure: [Second Slow System Model] Updated on 2024-03-26

[0111] ├─Applicable scenarios: 0.2pu≤U pcc <0.9pu;

[0112] ├─Core Function: Reactive Power Compensation Calculation;

[0113] └─Verification Case: Successfully supported voltage recovery to 0.85pu.

[0114] Therefore, a reduced-order model of the first slow system and a reduced-order model of the second slow system are collected, and the outputs of the two models are unified to the same time coordinate system. Parameters such as power and voltage are converted into per-unit values ​​(pu), and the reduced-order models of the first and second slow systems are dynamically coupled. The repetition range is determined based on the dynamic coupling of the reduced-order models of the first and second slow systems, and the repetitive low-voltage phenomenon is determined based on the analysis of the repetition range to mark the grid connection point of the photovoltaic grid-connected system. This approach is compatible with the overall consideration of the dynamic coupling of the reduced-order models of the first and second slow systems, ensuring the accuracy of the repetition range. At the same time, a slow system suitable for analyzing the normal steady-state control of the system is introduced, realizing the dynamic coupling of the reduced-order models of the first and second slow systems, ensuring the accuracy of the identification of the repetitive low-voltage phenomenon, and improving the accuracy of the identification of the grid connection point of the photovoltaic grid-connected system.

[0115] In the embodiments of this application, a preset repeated low-altitude penetration determination matching table is collected, as shown in Table 3:

[0116] Table 3. Matching Table for Repeated Low-Crossing Judgment

[0117] Parameters / conditions Normal range Repeated low-voltage trigger threshold Association Model Power deviation (ΔP) <10% of the rated value ≥15% for 5 seconds First vs. Second Model Output Voltage fluctuation amplitude <0.03% The price fluctuated repeatedly between 0.85 and 0.92 PU. PCC point measured data Event interval >30 minutes Repeated within 10 minutes Historical Event Records Reactive power compensation response delay <100ms >500ms Second Model Execution Log

[0118] Assuming an event analysis of a photovoltaic power station on March 28, 2024, the input parameters are: 1. Power deviation: Model 1 1.0 pu vs Model 2 0.7 pu (ΔP = 30%); 2. Voltage fluctuation: 0.87 0.89 pu for 8 seconds; 3. Recent events: 3 times at the same location within 24 hours; Based on the analysis of the repeated low-voltage test matching table, 1. ΔP = 30% > 15% √; 2. Voltage oscillation 0.87~0.89 pu √; 3. Repeated 3 times within 10 minutes √; Judgment output; "Confirmed repeated low-voltage test" → Mark grid connection point PCC4 as a yellow warning.

[0119] In step S14, the voltage dynamics of the photovoltaic grid-connected system in the short term after the reduction of active power output and the voltage dynamics in the short term after reactive power compensation are collected.

[0120] In the specific implementation of this invention, the specific steps are as follows:

[0121] S141: Collect data from the photovoltaic grid-connected system and mark the active power output stage of the photovoltaic grid-connected system. For the detection of the active power output stage of the photovoltaic grid-connected system, the first stage data corresponding to the detection is used to determine the voltage change curve of the photovoltaic grid-connected system in the short term after the active power output decreases based on the synthesis of the first stage data. Based on the voltage change curve of the photovoltaic grid-connected system in the short term after the active power output decreases, the voltage dynamics of the photovoltaic grid-connected system in the short term after the active power output decreases are determined.

[0122] S142: Collect data from the photovoltaic grid-connected system and mark the reactive power compensation stage of the photovoltaic grid-connected system. For the detection of the reactive power compensation stage of the photovoltaic grid-connected system, the second stage data corresponding to the detection is used to determine the voltage change curve of the photovoltaic grid-connected system in the short term after reactive power compensation based on the synthesis of the second stage data. The voltage dynamics of the photovoltaic grid-connected system in the short term after reactive power compensation are determined based on the voltage change curve of the photovoltaic grid-connected system in the short term after reactive power compensation.

[0123] In the embodiments of this application, data is collected from the photovoltaic grid-connected system to obtain relevant data during the phase of reduced active power output, in order to analyze voltage dynamic changes and mark the phase of reduced active power output by real-time monitoring. For example, when the output power of the photovoltaic system suddenly drops due to cloud cover or load changes, the start and end times of this phase are recorded.

[0124] The first phase of data includes active power, voltage, and current. This data needs to be collected in real-time during the active power output reduction phase; a sampling frequency of at least 100Hz is recommended to ensure the capture of transient voltage changes. The collected first-phase data will be synthesized into a time series to generate a voltage-time curve. Tools such as MATLAB and Python can be used for data processing and curve generation. By analyzing the voltage change curve, the trend and magnitude of voltage changes in the short term (e.g., within 100ms) after the active power output reduction are determined, and indicators such as the maximum voltage drop and recovery time are recorded.

[0125] Optionally, suppose that at a certain moment, the active power output of the photovoltaic system suddenly drops from 100kW to 50kW due to cloud cover. The following data is collected through a real-time monitoring system:

[0126] Time: 0.000s, Active power: 100kW, Voltage: 230V; Time: 0.050s, Active power: 90kW, Voltage: 228V; Time: 0.100s, Active power: 70kW, Voltage: 225V; Time: 0.150s, Active power: 50kW, Voltage: 220V; Time: 0.200s, Active power: 50kW, Voltage: 222V (Starting recovery);

[0127] Using this data, a voltage change curve over time was generated. It can be seen that the voltage drops rapidly after the active power output decreases, and then gradually recovers. Analysis of the curve shows that the maximum voltage drop is 10V (from 230V to 220V), and the recovery time is 0.1s.

[0128] Furthermore, relevant data from the system during the reactive power compensation phase is acquired to analyze voltage dynamic changes. By monitoring the system in real time, the stages of reactive power compensation are marked. For example, when the system detects a voltage drop and activates the reactive power compensation device (such as SVG or SVC), the start and end times of this stage are recorded.

[0129] The second stage of data includes reactive power, voltage, and current. This data needs to be collected in real-time during the reactive power compensation phase, with a recommended acquisition frequency of at least 100Hz to ensure the capture of transient voltage changes. The collected second-stage data is then used to synthesize a time series data set, generating a voltage-time curve. Tools such as MATLAB and Python can be used for data processing and curve generation.

[0130] By analyzing the voltage change curve, the trend and magnitude of voltage change in the short term (e.g., within 100ms) after reactive power compensation are determined, and indicators such as the maximum voltage rise and recovery time are recorded.

[0131] Assume that after the active power output decreases, the system activates the reactive power compensation equipment, increasing the reactive power from 0 kVar to 20 kVar. The following data was collected through the real-time monitoring system:

[0132] Time: 0.200s, reactive power: 0kVar, voltage: 220V; Time: 0.250s, reactive power: 10kVar, voltage: 223V; Time: 0.300s, reactive power: 20kVar, voltage: 228V; Time: 0.350s, reactive power: 20kVar, voltage: 230V (recovering to normal level); Using these data, a voltage change curve over time is generated. It can be seen that the voltage rises rapidly after reactive power compensation and then returns to normal. Analysis of the curve shows that the maximum voltage increase is 8V (from 220V to 228V), and the recovery time is 0.15s.

[0133] In step S15, effective control measures for the photovoltaic grid-connected system are determined based on the comparison between the voltage dynamics in the short term after the reduction of active power output and the voltage dynamics in the short term after reactive power compensation. The effective control measures are reactive power compensation triggered by the grid connection point.

[0134] In the specific implementation of this invention, the specific steps are as follows:

[0135] S151: Collect the voltage dynamics in the short term after the reduction of active power output and the voltage dynamics in the short term after reactive power compensation, compare the voltage dynamics in the short term after the reduction of active power output and the voltage dynamics in the short term after reactive power compensation, and determine the corresponding dynamic difference area based on the comparison of the voltage dynamics in the short term after the reduction of active power output and the voltage dynamics in the short term after reactive power compensation.

[0136] S152: Determine the corresponding difference content based on the detection of the dynamic difference area, and determine the effect coefficient of voltage dynamics in the short term after the reduction of active power output and the effect coefficient of voltage dynamics in the short term after reactive power compensation based on the identification of the difference content.

[0137] S153: The effective control measures for the photovoltaic grid-connected system are determined by comparing the effect coefficient of voltage dynamics in the short term after the reduction of active power output with the effect coefficient of voltage dynamics in the short term after reactive power compensation. The effective control measures are reactive power compensation triggered by the grid connection point.

[0138] In the embodiments of this application, voltage dynamic data is collected. Voltage dynamics after reduced active power output: voltage change curve and related data obtained in step S141; voltage dynamics after reactive power compensation: voltage change curve and related data obtained in step S142. By comparing the voltage dynamics under the two conditions, the effectiveness of reactive power compensation for voltage recovery is determined.

[0139] At this point, the voltage dynamic curve after the active power output is reduced is compared with the voltage dynamic curve after reactive power compensation. The voltage change trend and amplitude differences are analyzed to identify time periods or voltage ranges with significantly different voltage dynamic changes.

[0140] Optionally, assume that after a reduction in active power output, the voltage drops from 230V to 220V with a recovery time of 0.1s; while after reactive power compensation, the voltage rises from 220V to 230V with a recovery time of 0.15s. Comparing the two voltage dynamic curves: after a reduction in active power output: the voltage drops by 10V with a recovery time of 0.1s; after reactive power compensation: the voltage rises by 8V with a recovery time of 0.15s. The dynamic difference lies within 0.1s after the voltage drop and within 0.15s after reactive power compensation.

[0141] Furthermore, by analyzing the dynamic difference region, the specific impacts of reduced active power output and reactive power compensation on voltage dynamics are determined. Differences in key indicators such as voltage change amplitude and recovery time are calculated, including variations in voltage change amplitude and recovery time. Effectiveness coefficients for reduced active power output and reactive power compensation are introduced. The effectiveness coefficient for reduced active power output indicates the degree of impact of reduced active power output on voltage dynamics, while the effectiveness coefficient for reactive power compensation indicates the recovery effect of reactive power compensation on voltage dynamics.

[0142] Therefore, effective control measures for the photovoltaic grid-connected system are determined by comparing the effect coefficients of voltage dynamics in the short term after a reduction in active power output and the effect coefficients of voltage dynamics in the short term after reactive power compensation. The effective control measure is reactive power compensation triggered by the grid connection point. This comprehensive consideration takes into account both the effect coefficients of voltage dynamics in the short term after a reduction in active power output and the effect coefficients of voltage dynamics in the short term after reactive power compensation, ensuring the accuracy of the effective control measures for the photovoltaic grid-connected system. At the same time, it realizes the comparison of voltage dynamics in the short term after a reduction in active power output and the short term after reactive power compensation, improving the accuracy of the effective control measures for the photovoltaic grid-connected system and introducing an effective measure of reactive power compensation triggered by the grid connection point.

[0143] At this point, comparing the effect coefficient after the reduction in active power output with the effect coefficient after reactive power compensation, if the effect coefficient after reactive power compensation is significantly higher than that after the reduction in active power output, it indicates that reactive power compensation has a significant effect on voltage recovery. An effective control measure is reactive power compensation triggered at the grid connection point. When a voltage drop is detected, reactive power compensation equipment (such as SVG or SVC) is triggered at the grid connection point to quickly provide reactive power support and help restore the voltage. Simultaneously, when a voltage drop is detected, the reactive power compensation equipment is triggered at the grid connection point to quickly provide reactive power support and help the voltage return to normal levels.

[0144] In another embodiment of this application, the WECC electromechanical transient photovoltaic model will be used to construct a synchronous generator model and a model of the photovoltaic unit during normal operation, low voltage ride-through, and low voltage recovery.

[0145] The constructed synchronous motor model is as follows:

[0146]

[0147] Wherein, the superscript "." indicates the rate of change of the variable with respect to time (the same meaning applies below); δ represents the power angle; ω represents the rotational speed; ω0 represents the synchronous rotational speed; ω ref Indicates the reference rotational speed; T J P represents the inertial time constant; m P represents the input mechanical power; e Indicates electromagnetic power; D represents the damping coefficient; T represents electromagnetic power. d0 E′ represents the d-axis open-circuit transient time constant. q E represents the q-axis transient potential. fd Indicates the excitation voltage; X d Indicates the d-axis synchronous reactance; X′ d I represents the d-axis transient reactance; Gd T represents the d-axis current of the synchronous generator; A K represents the time constant of the excitation regulator. A V represents the gain coefficient of the excitation regulator; ref This indicates the internal reference voltage of the excitation regulator; V G Indicates the voltage at the generator terminal bus; I Gq X represents the q-axis current of the synchronous generator; q Indicates the q-axis synchronous reactance; V Gd and V Gq These represent the d-axis and q-axis voltages of the synchronous generator, respectively.

[0148] The dynamic equations during normal operation of the photovoltaic unit are as follows:

[0149] Active current control section:

[0150]

[0151] Among them, T pord P represents the photovoltaic active power command time constant; ord Indicates photovoltaic active power command; P ref V represents the reference value for photovoltaic active power. PV I′ represents the photovoltaic node voltage in the xyz coordinate system. pcmd Indicates the intermediate variable for photovoltaic active current command; I max Indicates the maximum current limit of photovoltaic equipment; I qcmd Indicates photovoltaic reactive current command; I pcmd Indicates photovoltaic active current command; I pmax Indicates the maximum photovoltaic active power control current; I pmin T represents the minimum photovoltaic active power control current; g This represents the lag time constant of the converter current regulator.

[0152] Reactive current control section:

[0153]

[0154] Among them, T iq Indicates the photovoltaic reactive power command time constant; I qext Indicates photovoltaic reactive current command; Q ext Indicates the reference value for photovoltaic reactive power; I′ qcmd Indicates the intermediate variable for photovoltaic reactive current command; I qmax Indicates the maximum photovoltaic reactive power control current; I qmin This represents the minimum photovoltaic reactive power control current.

[0155] The dynamic equations for the low-voltage ride-through of a photovoltaic (PV) unit can be expressed as follows:

[0156] Active current control section:

[0157]

[0158] Reactive current control section:

[0159]

[0160] Among them, I qinj K represents the injected reactive current during the photovoltaic low-voltage ride-through period, which is the dynamic voltage support. qv V represents the photovoltaic reactive current injection gain; ref0 This represents the reference voltage for reactive current injection, typically taken as 0.9 pu.

[0161] The dynamic equations for the low-voltage ride-through recovery period of a photovoltaic (PV) unit can be expressed as:

[0162] Active current control section:

[0163]

[0164] Among them, I prmax Indicates the rate of rise limit of the active current command; K rrpwr Indicates the active current recovery rate; Δt represents the time interval from the initial moment of the low-voltage ride-through recovery phase; I p0 y represents the active current at the initial moment of the low-voltage ride-through recovery phase; y represents the intermediate variable of the photovoltaic active control current.

[0165] Reactive current control section:

[0166]

[0167] The network equations of a power system can be expressed as:

[0168]

[0169] Where Y′ represents the node admittance matrix that has been incorporated into the transient reactance of the synchronous generator and the constant impedance load; I G =E′ q . / (jX′ d ) represents the Norton equivalent injection current of the synchronous generator; S PV S represents the photovoltaic injected apparent power vector; L This represents the load power vector.

[0170] The interface equation of a photovoltaic unit can be expressed as:

[0171]

[0172] Among them, P PV Q represents the active power injected by photovoltaics; PV This indicates the reactive power injected by the photovoltaic system.

[0173] The i-th load interface equation can be expressed as (subscript i omitted):

[0174] S L =S LCP

[0175] Among them, S LCP This indicates a constant power load.

[0176] Step 102: Determine whether the system can be decomposed based on Tikhonov's fundamental theorem.

[0177] Consider the following dual-timescaled system:

[0178]

[0179] Where ε is the singular perturbation parameter, satisfying 0 < ε << 1; x F x represents the system's fast state variable vector; S This represents the system's slow state variable vector.

[0180] If vector functions f1 and f2 are differentiable in higher orders, and the system satisfies the following conditions:

[0181] 1) For a given initial value x of a slow system S0 The equilibrium point of the fast system is asymptotically stable, and the initial value of the fast variable x is... F0 It is located within the attraction domain of the fast system.

[0182] 2) On the solution curve of the slow system, the Jacobian matrix The real part of the eigenvalues ​​is strictly less than 0, that is:

[0183]

[0184] The solutions of the fast system and the slow system can consistently approximate the solution of the original system, and the original dual-timescaled system can be decomposed into a fast system and a slow system.

[0185] Fast System:

[0186]

[0187] Slow systems:

[0188]

[0189] Given T′ d0 The order of magnitude is generally 10 0 ~10 1 And T A T Pord T g T iq The order of magnitude is generally 10 -2 ~10 -1 Now consider the system's dual-timescale decomposition during normal photovoltaic operation.

[0190] Considering the normal operation of the photovoltaic system, E′ can be assumed to be... q E is a slow state variable. fd P ord I p I qext and I q For fast state variables, that is:

[0191]

[0192] During normal operation of a photovoltaic system, the reactive power output is approximately zero, and both active and reactive currents will not exceed the limits. Therefore, a fast system can be obtained.

[0193]

[0194] The Jacobian matrix can be obtained. The expression:

[0195]

[0196] in:

[0197]

[0198]

[0199] The order of magnitude of grid voltage-active power sensitivity and voltage-reactive power sensitivity is typically 10. -2 ~10 -8The sensitivity at the output of the synchronous generator is approximately 0; T A The value range of K is generally 0.05 to 0.1; A The value range of T is generally 20 to 200; Pord T iq The value range of T is generally 0.01 to 0.1; g The value range is generally 0.017 to 0.05; when 100MVA is taken as the reference voltage, P ord With Q ext The value range is generally 10. -2 ~10 0 Voltage is taken as a per-unit value.

[0200] It can be seen that the Jacobian matrix is ​​a block upper triangular matrix, and its eigenvalues ​​are equal to the union of the eigenvalues ​​obtained from the two diagonal block matrices. It is also easy to see that the upper left block is a diagonal matrix, and the lower right block is approximately a weakly strictly diagonally dominant matrix. Therefore, the real parts of all eigenvalues ​​of the Jacobian matrix are less than 0, satisfying condition (2) of the fundamental theorem of Tikhonov.

[0201] Condition (1) can be verified by simulation using the IEEE 3-machine 9-node system example. The specific relevant parameters of the synchronous generator and photovoltaic unit are shown in Table 1:

[0202] Table 1. Relevant parameters of synchronous generators and photovoltaic units in IEEE 3-machine 9-node system

[0203] parameter <![CDATA[K A ]]> <![CDATA[T A ]]> <![CDATA[T Pord ]]> <![CDATA[T iq ]]> <![CDATA[T g ]]> <![CDATA[X′ d ]]> value 175 0.1 0.1 0.016668 0.02 0.0608

[0204] like Figure 2 As shown, in the IEEE 3-machine 9-node example, the real parts of the Jacobian matrix eigenvalues ​​of the fast system under normal operating conditions are all less than 0, satisfying the Tikhonov fundamental theorem.

[0205] Step 103: Use singular perturbation theory to perform dual time-scale decomposition on the model to obtain a fast system suitable for analyzing the dynamic characteristics of the system and a slow system suitable for analyzing the normal steady-state control of the system.

[0206] Here is a slow system under normal operating conditions:

[0207]

[0208] It is easy to see that the slow system model has a decisive influence on the movement of the equilibrium point of the original system during normal photovoltaic operation.

[0209] Similarly, a dual-timescale decomposition can be performed on the system considering both the photovoltaic low-voltage ride-through period and the photovoltaic low-voltage ride-through recovery period. Here, the dual-timescale decomposition results of the system during the photovoltaic low-voltage ride-through period are given.

[0210] During the photovoltaic low-voltage ride-through period, the photovoltaic reactive current reaches I max System dual-timescale decomposition of the scenario:

[0211] Fast System:

[0212]

[0213] Slow systems:

[0214]

[0215] During the low-voltage ride-through period, the photovoltaic reactive current did not reach I. max The system's dual-timescale decomposition in the scenario: Fast system:

[0216]

[0217] Slow systems:

[0218]

[0219] Step 104: Based on the slow system reduced-order model during normal photovoltaic operation and low-voltage ride-through, a mathematical explanation is given for the mechanism by which repeated low-voltage ride-through easily occurs in areas with a high proportion of photovoltaic power.

[0220] When the active power of a photovoltaic unit connected to a weak grid increases, repeated low-voltage ride-through may occur. Its switching characteristics are a process in which the photovoltaic unit repeatedly switches between two stages: normal operation and low-voltage ride-through.

[0221] In the IEEE 3-machine 9-node example, at 0.5s, the output of the photovoltaic units located at nodes 2 and 3 increases by 10MW respectively. The voltage waveforms and operating status change curves of the photovoltaic units at the grid connection point during repeated low-voltage runs are shown below. Figure 3 and Figure 4 As shown. When the voltage at node 2 drops, the photovoltaic unit connected to node 2 enters repeated low voltage ride-through at 0.77s. During the repeated low voltage ride-through, the photovoltaic unit only switches between state 1 and state 2 repeatedly and does not enter state 3 (where 1 indicates that the photovoltaic is in the normal operation stage, 2 indicates that the photovoltaic is in the low voltage ride-through stage, and 3 indicates that the photovoltaic is in the low voltage ride-through recovery stage).

[0222] Simulation verification was performed in case 1354 pegase. The specific relevant parameters of the synchronous generator and photovoltaic unit are shown in Table 2.

[0223] Table 2. Relevant parameters of synchronous generators and photovoltaic units in IEEE 3-machine 9-bus system

[0224] parameter <![CDATA[K A ]]> <![CDATA[T A ]]> <![CDATA[T Pord ]]> <![CDATA[T iq ]]> <![CDATA[T g ]]> <![CDATA[X′ d ]]> value 175 0.1 0.1 0.016668 0.02 0.0608

[0225] In the case 1354 pegase, at 0.5s, the output of the photovoltaic units located at nodes 1794 and 7808 increased by 580MW respectively. The voltage waveforms and operating status change curves of the photovoltaic units at the grid connection point during repeated low-voltage runs are shown below. Figure 5 and Figure 6 As shown, the photovoltaic unit connected to node 1794 entered repeated low-voltage testing at 0.96s. During the repeated low-voltage testing, the photovoltaic unit only switched between state 1 and state 2 repeatedly, and did not enter state 3.

[0226] As can be seen from the slow system during normal operation of photovoltaic units, when the photovoltaic active power generation increases, P ref An increase in P PV Increase.

[0227] Since the system voltage equilibrium point satisfies the network equations, the network equations are analyzed.

[0228] Consider a power grid containing synchronous generators, photovoltaic systems, constant-power loads, and constant-impedance loads, where the transient potentials of the constant-impedance loads and synchronous generators are incorporated into the grid node admittance matrix. Its network equations can be written in the following form:

[0229]

[0230] When analyzing the repetitive low-voltage problem, since the node voltage variation range is small, around 0.9 pu, the following formula is satisfied:

[0231]

[0232] Where V0 represents the initial value of the node voltage; S represents the changed power value.

[0233] Let S = S PV -S L The network equation can be rewritten as:

[0234]

[0235] Considering that repeated low-voltage transmission is a local problem, when photovoltaic output increases in a high-PV-penetration area of ​​the grid, it only affects the voltage of that local area, while the voltage of distant nodes can be considered essentially unchanged. Therefore, the network equations can be rewritten as follows:

[0236]

[0237] In this context, the subscript "N" represents a nearby node; the subscript "F" represents a distant node.

[0238] Furthermore, we can obtain:

[0239]

[0240] When the voltage 0° reference phase is set in this local area, we can obtain:

[0241]

[0242] During the normal operation of the photovoltaic unit, due to Y′ NN The order of magnitude is much larger than Therefore, the above equation can be approximated as:

[0243]

[0244] Finally, it is easy to see from the slow system that the new system's E′ after the photovoltaic active power output increases. q The change is minimal, remaining around 1.0 pu. Meanwhile, X′ d The order of magnitude is generally 10 -2 ~10 -1 Therefore, it can be concluded that during the normal operation phase, due to I GN The values ​​are all much greater than Therefore, for grids with low synchronous generator operation and high photovoltaic penetration, the voltage balance point drop caused by the increase in photovoltaic output is greater than that of grids with high synchronous generator operation, and is more likely to drop below 0.9 pu, becoming an unstable balance point.

[0245] Step 105: Based on the theory of monotonic dynamic systems with input and output, compare two emergency power voltage control measures: reducing the active power output of photovoltaic power and performing reactive power compensation at the photovoltaic output, and obtain the optimal control strategy.

[0246] Consider a nonlinear controlled system with output:

[0247] Σ: y = k(x,v)

[0248] Where x represents the state variable vector; v represents the control variable vector; and y represents the output variable vector.

[0249] The system ∑ has the property of monotonicity and order preservation of input-output, and the theorem is given below.

[0250] definition:

[0251] Υ(σ)=y(t,φ2+σ(φ1-φ2),v2+σ(v1-v2))

[0252] but:

[0253]

[0254] in:

[0255]

[0256] For v1~v2 and t≥0, if Υ′(σ)≥0, the following input-output monotonic order-preserving property can be obtained:

[0257] Δy=y(t,φ1,v1)-y(t,φ2,v2)≥0

[0258] Conversely, if Υ′(σ)≤0, then:

[0259] Δy=y(t,φ1,v1)-y(t,φ2,v2)≤0

[0260] Where φ and y(t,φ,v) represent the vectors of solutions to x and y, respectively; t represents time; v1~v2 means that each element in v1 is greater than or equal to the corresponding element in v2; Δy represents the vector of changes in the output variable.

[0261] Considering the short-term voltage dynamics after the reduction in active power output of the photovoltaic unit, the following can be obtained from the fast system electromechanical transient model of the photovoltaic grid-connected system during normal operation:

[0262]

[0263] Among them, |V lv | represents the amplitude vector of the photovoltaic grid connection point voltage that repeatedly experiences low voltage drops during normal operation of the photovoltaic unit.

[0264] because and The reactive current of the photovoltaic system remains constant at 0 during normal operation. (0); in high-proportion photovoltaic regional power grids Generally, it is negative. Therefore, we can conclude that:

[0265]

[0266] Among them, I p1 This represents the photovoltaic active current vector before the reduction in photovoltaic active power output during normal operation; I p2 This represents the photovoltaic active current vector after the photovoltaic active power output decreases during normal operation; |V lv1 | represents the vector of photovoltaic output voltage amplitude before changes in photovoltaic active power during normal operation; |V lv2 | represents the photovoltaic output voltage amplitude vector after the change in photovoltaic active power during normal operation.

[0267] Considering the voltage dynamics near the equilibrium point after the active power output of the photovoltaic unit decreases, the electromechanical transient model of the slow system including the photovoltaic grid-connected system during normal operation can be obtained as follows:

[0268]

[0269] Because of E′ throughout the dynamic process q The change is minimal, remaining around 1.0 pu. Therefore, during normal operation, the decrease in photovoltaic active power output can raise the photovoltaic output voltage amplitude, avoiding repeated low-voltage breakdown. The voltage rise is from Y′ P (σ) determines.

[0270] Considering the short-term voltage dynamics after reactive power compensation, the electromechanical transient model of a grid-connected photovoltaic system during normal photovoltaic operation yields the following:

[0271]

[0272] Among them, Q C This represents the reactive power compensation vector at the photovoltaic outlet where repeated low-voltage driving occurs.

[0273] Since the order of magnitude of the grid voltage-reactive power sensitivity is much larger than that of the voltage-active power sensitivity, we can conclude that:

[0274]

[0275] Considering the voltage dynamics near the equilibrium point after reactive power compensation, the electromechanical transient model of a grid-connected photovoltaic system during normal operation yields the following:

[0276]

[0277] Because of E′ throughout the dynamic process q The change is minimal, remaining around 1.0 pu, thus the aforementioned conclusion still holds. Therefore, reactive power compensation at the photovoltaic output during normal operation can increase the photovoltaic output voltage amplitude, avoiding repeated low-voltage breakdowns. The voltage rise is reduced from Y′ Q (σ) determines this. It is easy to see that when the power regulation value is the same, Υ′ Q (σ) is much larger than Υ′ P Therefore, reactive power compensation at the photovoltaic outlet during normal operation is a more effective control measure.

[0278] Please see Figure 2 , Figure 2 This is a schematic diagram of the structural composition of the control system of the photovoltaic grid-connected system based on the low-voltage ride-through scenario in an embodiment of the present invention; the control system of the photovoltaic grid-connected system based on the low-voltage ride-through scenario includes:

[0279] The partitioning module 21 is used to collect the electromechanical transient control model of the photovoltaic grid-connected system, and to determine the normal operation period, low voltage ride-through period and low voltage ride-through recovery period of the photovoltaic grid-connected system based on the partitioning of the electromechanical transient control model.

[0280] The decomposition module 22 is used to perform dual time-scale decomposition on the electromechanical transient control model of the photovoltaic grid-connected system during normal operation, so as to output a fast system suitable for analyzing the dynamic characteristics of the system and a slow system suitable for analyzing the normal steady-state control of the system.

[0281] The grid connection point module 23 is used to collect the first slow system reduced-order model during normal operation and the second slow system reduced-order model during low voltage ride-through in a slow system. Based on the dynamic coupling of the first slow system reduced-order model and the second slow system reduced-order model, the repeated low voltage ride-through phenomenon is determined, and the grid connection point of the photovoltaic grid-connected system is marked.

[0282] Voltage dynamic module 24 is used to collect voltage dynamics of the photovoltaic grid-connected system in the short term after the active power output decreases and in the short term after reactive power compensation.

[0283] The effective control measures module 25 is used to determine the effective control measures of the photovoltaic grid-connected system based on the comparison of the voltage dynamics in the short term after the reduction of active power output and the voltage dynamics in the short term after reactive power compensation. The effective control measures are reactive power compensation triggered by the grid connection point.

[0284] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

Claims

1. A control method for a photovoltaic grid-connected system based on a low-voltage ride-through scenario, characterized in that, include: The electromechanical transient control model of the photovoltaic grid-connected system is collected, and the normal operation period, low voltage ride-through period, and low voltage ride-through recovery period of the photovoltaic grid-connected system are determined based on the division of the electromechanical transient control model. During the normal operation of the photovoltaic grid-connected system, the electromechanical transient control model of the photovoltaic grid-connected system is decomposed into a dual time scale to output a fast system suitable for analyzing the dynamic characteristics of the system and a slow system suitable for analyzing the normal steady-state control of the system. In the slow system, the first slow system reduced-order model during normal operation and the second slow system reduced-order model during low voltage ride-through are collected. Based on the dynamic coupling of the first and second slow system reduced-order models, the repeated low voltage ride-through phenomenon is determined, and the grid connection point of the photovoltaic grid-connected system is marked. Collect voltage dynamics of the photovoltaic grid-connected system in the short term after a reduction in active power output and in the short term after reactive power compensation; Effective control measures for photovoltaic grid-connected systems are determined by comparing the short-term voltage dynamics after a reduction in active power output with the short-term voltage dynamics after reactive power compensation. The effective control measure is reactive power compensation triggered by the grid connection point.

2. The control method for a photovoltaic grid-connected system based on a low-voltage ride-through scenario according to claim 1, characterized in that, The electromechanical transient control model of the photovoltaic grid-connected system is used to determine the normal operation period, low voltage ride-through period, and low voltage ride-through recovery period of the photovoltaic grid-connected system, based on the division of the electromechanical transient control model. Real-time monitoring of the photovoltaic grid-connected system, collection of multiple operating data of the photovoltaic grid-connected system, and determination of the electromechanical transient control model of the photovoltaic grid-connected system based on the multiple operating data of the photovoltaic grid-connected system and the corresponding control framework logic of the photovoltaic grid-connected system; Multiple working paths are determined based on the detection of the electromechanical transient control model of the photovoltaic grid-connected system. In each working path, the path nodes of each working path are marked, and the corresponding working data is matched for each path node. The period type corresponding to the working path is determined based on the location of each path node and the working data corresponding to each path node. This period type includes normal operation period, low voltage ride-through period, and low voltage ride-through recovery period.

3. The control method for a photovoltaic grid-connected system based on a low-voltage ride-through scenario according to claim 1, characterized in that, During the normal operation of the photovoltaic grid-connected system, the electromechanical transient control model of the photovoltaic grid-connected system is decomposed into a dual-time-scale model to output a fast system suitable for analyzing the dynamic characteristics of the system and a slow system suitable for analyzing the normal steady-state control of the system, including: During the normal operation of the photovoltaic grid-connected system, the electromechanical transient control model of the photovoltaic grid-connected system is judged based on the Tikhonov fundamental theorem in order to define whether it conforms to the range of dual time-scale decomposition. If the range of dual-time-scale decomposition is met, the dual-time-scale decomposition of the electromechanical transient control model of the photovoltaic grid-connected system is triggered, and the corresponding fast system and slow system are output. At this time, the fast system is suitable for analyzing the dynamic characteristics of the system; the slow system is suitable for analyzing the normal steady-state control of the system.

4. The control method for a photovoltaic grid-connected system based on a low-voltage ride-through scenario according to claim 1, characterized in that, In the slow system, a first slow system reduced-order model during normal operation and a second slow system reduced-order model during low-voltage ride-through are collected. Based on the dynamic coupling of the first and second slow system reduced-order models, repeated low-voltage ride-through phenomena are determined, and the grid connection point of the photovoltaic grid-connected system is marked, including: Slow systems are monitored in real time, and the range of slow systems during different periods is marked to determine multiple first system data during normal operation and multiple second system data during low voltage ride-through. During normal operation, the first slow system reduced-order model is determined based on multiple first system data and the expression corresponding to the slow system; during low voltage ride-through, the second slow system reduced-order model is determined based on multiple second system data and the expression corresponding to the slow system.

5. The control method for a photovoltaic grid-connected system based on a low-voltage ride-through scenario according to claim 4, characterized in that, In the slow system, the method involves collecting a first slow system reduced-order model during normal operation and a second slow system reduced-order model during low-voltage ride-through. Based on the dynamic coupling of the first and second slow system reduced-order models, repeated low-voltage ride-through phenomena are determined, and the grid connection point of the photovoltaic grid-connected system is marked. The method also includes: The first slow system reduced-order model and the second slow system reduced-order model are collected. The first slow system reduced-order model and the second slow system reduced-order model are dynamically coupled. The repetition range is determined based on the dynamic coupling of the first slow system reduced-order model and the second slow system reduced-order model. The repetition low-pass phenomenon is determined based on the analysis of the repetition range, so as to mark the grid connection point of the photovoltaic grid-connected system.

6. The control method for a photovoltaic grid-connected system based on a low-voltage ride-through scenario according to claim 1, characterized in that, The voltage dynamics of the photovoltaic grid-connected system in the short term after a reduction in active power output and in the short term after reactive power compensation include: Data is collected from the photovoltaic grid-connected system, and the active power output stage of the photovoltaic grid-connected system is marked. The first stage data corresponding to the detection of the active power output stage of the photovoltaic grid-connected system is used to determine the voltage change curve of the photovoltaic grid-connected system in the short term after the active power output decreases, based on the synthesis of the first stage data. The voltage dynamics of the photovoltaic grid-connected system in the short term after the active power output decreases are determined based on the voltage change curve of the photovoltaic grid-connected system in the short term after the active power output decreases.

7. The control method for a photovoltaic grid-connected system based on a low-voltage ride-through scenario according to claim 6, characterized in that, The method for collecting voltage dynamics of the photovoltaic grid-connected system in the short term after a reduction in active power output and in the short term after reactive power compensation also includes: Data is collected from the photovoltaic grid-connected system, and the reactive power compensation stage of the photovoltaic grid-connected system is marked. The second stage data corresponding to the detection of the reactive power compensation stage of the photovoltaic grid-connected system is used to determine the voltage change curve of the photovoltaic grid-connected system in the short term after reactive power compensation based on the synthesis of the second stage data. The voltage dynamics of the photovoltaic grid-connected system in the short term after reactive power compensation are determined based on the voltage change curve of the photovoltaic grid-connected system in the short term after reactive power compensation.

8. The control method for a photovoltaic grid-connected system based on a low-voltage ride-through scenario according to claim 1, characterized in that, The effective control measures for the photovoltaic grid-connected system are determined by comparing the short-term voltage dynamics after a reduction in active power output with the short-term voltage dynamics after reactive power compensation. These effective control measures include reactive power compensation triggered at the grid connection point, comprising: The voltage dynamics in the short term after the reduction of active power output and the short term after reactive power compensation are collected. The voltage dynamics in the short term after the reduction of active power output and the short term after reactive power compensation are compared. Based on the comparison of the voltage dynamics in the short term after the reduction of active power output and the short term after reactive power compensation, the corresponding dynamic difference area is determined.

9. The control method for a photovoltaic grid-connected system based on a low-voltage ride-through scenario according to claim 8, characterized in that, The effective control measures for the photovoltaic grid-connected system are determined by comparing the short-term voltage dynamics after a reduction in active power output with the short-term voltage dynamics after reactive power compensation. The effective control measures include reactive power compensation triggered at the grid connection point, and further include: Based on the detection of dynamic difference regions, the corresponding difference content is determined, and based on the identification of the difference content, the effect coefficient of voltage dynamics in the short term after the reduction of active power output and the effect coefficient of voltage dynamics in the short term after reactive power compensation are determined. Effective control measures for the photovoltaic grid-connected system are determined by comparing the effect coefficient of voltage dynamics in the short term after the reduction of active power output with the effect coefficient of voltage dynamics in the short term after reactive power compensation. The effective control measure is reactive power compensation triggered by the grid connection point.

10. A control system for a photovoltaic grid-connected system based on a low-voltage ride-through scenario, characterized in that, The control system of the photovoltaic grid-connected system based on the low-voltage ride-through scenario is applied to the control method of the photovoltaic grid-connected system based on the low-voltage ride-through scenario as described in any one of claims 1-9, wherein the control system of the photovoltaic grid-connected system based on the low-voltage ride-through scenario includes: The partitioning module is used to collect the electromechanical transient control model of the photovoltaic grid-connected system, and to determine the normal operation period, low voltage ride-through period and low voltage ride-through recovery period of the photovoltaic grid-connected system based on the partitioning of the electromechanical transient control model. The decomposition module is used to perform dual-time-scale decomposition of the electromechanical transient control model of the photovoltaic grid-connected system during normal operation, so as to output a fast system suitable for analyzing the dynamic characteristics of the system and a slow system suitable for analyzing the normal steady-state control of the system. The grid connection point module is used to collect the first slow system reduced-order model during normal operation and the second slow system reduced-order model during low voltage ride-through in a slow system. Based on the dynamic coupling of the first and second slow system reduced-order models, the repeated low voltage ride-through phenomenon is determined, and the grid connection point of the photovoltaic grid-connected system is marked. The voltage dynamics module is used to collect the voltage dynamics of the photovoltaic grid-connected system in the short term after the active power output decreases and in the short term after reactive power compensation. The effective control measures module is used to determine the effective control measures for the photovoltaic grid-connected system based on the comparison of the voltage dynamics in the short term after the reduction of active power output and the voltage dynamics in the short term after reactive power compensation. The effective control measures are reactive power compensation triggered by the grid connection point.

Images