A photovoltaic inverter operation control system and method based on multi-source data
By building a three-loop architecture and optimizing the control of the photovoltaic inverter using an LSTM model, the problems of current surges and voltage fluctuations in traditional photovoltaic inverters are solved, achieving efficient power transmission and grid-friendliness.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SHANDONG HUADIAN ENERGY CONSERVATION TECHNOLOGY CO LTD
- Filing Date
- 2025-12-09
- Publication Date
- 2026-05-15
AI Technical Summary
In the operation and control of traditional photovoltaic inverters, there is a lack of predictive control of the closing current. The adjustment of the closing or opening rate is lagging, which leads to the spread of current surges and voltage fluctuations, increased transmission losses, and reduced power plant power generation efficiency.
A three-circuit architecture of main switching, bypass, and suppression is constructed. Power generation conditions are judged by combining multi-source data. The LSTM prediction model is used to optimize the closing rate of the vacuum contactor, select the optimal path, monitor the grid voltage in real time and dynamically adjust the closing rate, and suppress current surges by combining current limiting resistors.
It effectively avoids the impact of closing current, reduces transmission loss, improves the power generation efficiency of the power plant, ensures the stability of the power grid, and prevents the spread of voltage fluctuations.
Smart Images

Figure CN121395877B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of inverter control technology, specifically a photovoltaic inverter operation control system and method based on multi-source data. Background Technology
[0002] Photovoltaic power generation, as a core clean energy source, has achieved large-scale development. Inverters, as the core component connecting photovoltaic systems to the grid, directly determine the power generation efficiency and grid connection security of the power plant through their stable and efficient operation and control. Collector lines, as the key carriers for transmitting electrical energy in inverters, have their on / off control, path switching, and fault suppression as core aspects of inverter operation and control. During commissioning, it is necessary to avoid core saturation caused by closing current surges; during operation, it is necessary to reduce path transmission losses; and during shutdown, it is necessary to prevent voltage fluctuations from spreading to the grid.
[0003] Traditional commissioning relies solely on fixed closing rate control, failing to accurately assess low-load conditions on both the grid and inverter sides, and lacks predictive control of closing current. Furthermore, the timing of suppression circuit connection and resistor current limiting effectiveness are poorly matched. Bypass circuit selection often depends on manual presets or single distance parameters, neglecting the correlation between the photovoltaic array's main power generation area and circuit location. This easily leads to the selection of suboptimal paths, increasing transmission losses and reducing power plant efficiency. Monitoring of grid-connected voltage fluctuations during operation and shutdown is mostly reactive, lacking trend prediction and dynamic adjustment mechanisms. Lagging adjustments to closing or opening rates can easily cause fluctuations to spread to the grid. Summary of the Invention
[0004] The purpose of this invention is to provide a photovoltaic inverter operation control system and method based on multi-source data to solve the problems raised in the prior art.
[0005] To achieve the above objectives, the present invention provides the following technical solution:
[0006] In a first aspect, this application provides a photovoltaic inverter operation control method based on multi-source data, comprising the following steps:
[0007] In the collector lines associated with the inverter, main switching circuit, bypass circuit and suppression circuit are built; multi-source data are collected to determine whether the photovoltaic power station has the conditions for power generation and generate control commands, including commissioning commands and shutdown commands;
[0008] When the commissioning command is triggered, the detection equipment confirms whether the current on the grid side and the inverter side is in a low load state. If the current exceeds the standard, the commissioning is delayed. The LSTM prediction model is used to control the vacuum contactor of the main switching circuit to close, and the suppression circuit is connected to the circuit to slow down the current rise rate at the moment of closing by using a resistor.
[0009] After the main switching circuit current stabilizes, the optimal path is selected based on the orientation relationship between the main switching circuit and the bypass circuit and the photovoltaic power station, and the operation is switched to the bypass circuit to disconnect the suppression circuit; the grid voltage is monitored in real time, and the closing rate of the main switching circuit is adjusted when a sudden rise or fall in voltage occurs.
[0010] When the shutdown command is triggered, the main on / off circuit is opened, disconnecting the main connection between the collector line and the inverter, while the bypass circuit remains operational. After the main on / off circuit is opened, the current and voltage status of the bypass circuit are detected. Once it is confirmed that there is no abnormal overcurrent or overvoltage, the bypass circuit is disconnected. When voltage fluctuations exceed the limit, the interlocking is triggered, and the opening operation is suspended.
[0011] In conjunction with the first aspect, in a first embodiment of the first aspect of this application, the construction of a main switching circuit, a bypass circuit, and a suppression circuit in the collector lines associated with the inverter includes:
[0012] The main switching circuit is responsible for the main control switching of the collector line during commissioning and shutdown. The bypass circuit is responsible for the current diversion after the main switching circuit stabilizes. The suppression circuit is responsible for suppressing inrush current and voltage fluctuations during commissioning.
[0013] The main switching circuit is connected in series between the inverter output and the collector bus to form the main current path. A disconnect switch, a vacuum contactor, and a current transformer are connected in series within the circuit. Voltage monitoring points are connected in parallel across the circuit, and voltage transformers are connected to these points to monitor the voltage across the main switching circuit in real time. A bypass circuit is connected in parallel across the vacuum contactor of the main switching circuit, forming a parallel structure between the main switching circuit and the bypass circuit. A bypass disconnect switch, a bypass contactor, and a current transformer are configured within this bypass circuit, which is grouped according to the photovoltaic array orientation. A suppression circuit is connected in series between the vacuum contactor of the main switching circuit and the collector bus to form a series path. A current-limiting resistor and a suppression circuit contactor are configured within this circuit. Voltage monitoring points are connected in parallel across the suppression circuit to assist in determining the voltage division when the resistor is connected.
[0014] In conjunction with the first aspect, in the second embodiment of the first aspect of this application, the step of collecting multi-source data, determining whether the photovoltaic power station has the conditions for power generation, and generating control commands, including commissioning commands and shutdown commands, includes:
[0015] Multi-source data includes photovoltaic module operating condition data, grid connection status data, environmental interference data, and equipment health status data. Based on photovoltaic power plant design standards and equipment parameters, effective thresholds for each type of multi-source data are preset. Multi-source data is collected, pre-processed, and compared with the effective thresholds. When the module has power generation capacity, the grid has acceptance capacity, and the equipment has operational safety, a commissioning command is generated; otherwise, a shutdown command is generated.
[0016] In conjunction with the first aspect, in the third embodiment of the first aspect of this application, the step of confirming whether the current on the grid side and the inverter side is in a low-load state through detection equipment after the commissioning command is triggered, and delaying commissioning when the current exceeds the standard, includes:
[0017] Based on the inverter's rated parameters, collector line design standards, and inrush current suppression targets, a quantification threshold for low-load conditions is set. Pre-calibration is performed on the detection equipment on both the grid and inverter sides. Upon triggering the commissioning command, current acquisition on both the grid and inverter sides is initiated, and data preprocessing is performed. If both the grid-side and inverter-side currents are not greater than the low-load threshold, the system is determined to be in a low-load state; otherwise, the current is deemed to be exceeding the limit. If the current exceeds the limit, the main switching circuit closing action is frozen, and an initial delay time is set. After the delay time expires, the system re-evaluates whether the low-load requirements are met. If the current still exceeds the limit, the verification process is repeated until the low-current operating condition requirements are met.
[0018] In conjunction with the first aspect, in the fourth embodiment of the first aspect of this application, the method of using an LSTM predictive model to control the closing of the vacuum contactor in the main switching circuit, by connecting a suppression circuit to the circuit and using a resistor to slow down the rate of current rise at the moment of closing, includes:
[0019] Time-series data of past closing processes in photovoltaic power plants are collected to construct a training sample library. Input time-series features include grid-side voltage sequences, inverter-side initial current sequences, irradiance sequences, module temperature sequences, and real-time temperature sequences of suppression resistors. Output time-series labels, specifically the current rise sequence after closing. The data is cleaned and normalized, and divided into training and validation sets. An LSTM prediction model is constructed with the goal of minimizing the mean square error between the predicted current sequence and the actual sequence. The Adam optimizer is used for training to ensure that the prediction error is within the allowable range. The collected current time-series features are input into the LSTM model, and the prediction results, specifically the current rise curve, are output. Based on the curve, the predicted peak value, predicted rise slope, and predicted core saturation time are extracted to generate control parameters for the closing and suppression circuits. Based on the control parameters, a stepped closing command is sent to the vacuum contactor, and an early closing command is sent to the suppression circuit contactor.
[0020] In conjunction with the first aspect, in the fifth embodiment of the first aspect of this application, after the current in the main switching circuit stabilizes, the optimal path is selected based on the location relationship between the main switching circuit and the bypass circuit and the photovoltaic power station, and the system switches to bypass circuit operation, disconnecting the suppression circuit, including:
[0021] The system calls up pre-stored power plant topology data, which defines the physical connection relationships and electrical paths between all bypass circuits and each group of photovoltaic arrays. The data attributes include circuit identifier, associated array group, azimuth coordinates, and current status. The associated array group is a subset of photovoltaic arrays connected to the circuit. The azimuth coordinates are the relative electrical distance between the circuit access point and the array group. The current status is the status of the disconnector and contactor of the bypass circuit.
[0022] The selection principle is to choose the bypass circuit that is closest to the current main power generation array in terms of location. Real-time power generation data of each photovoltaic array group is obtained, and one or more array groups contributing the main power generation are identified. The identified main power generation array groups are matched with the array groups associated with each bypass circuit. The path priority score of each bypass circuit is calculated, with the scoring logic being: if the array group associated with a bypass circuit has the greatest overlap with the current main power generation array group, its score is the highest. The bypass circuit with the highest path priority score is selected as the target bypass circuit for this switchover. When multiple circuits with the same score appear, the one with the shortest electrical path is selected. An operation interlock signal is sent to the vacuum contactor of the main switching circuit to close the bypass isolating switch of the target bypass circuit. Subsequently, the bypass contactor of that circuit is closed. When it is confirmed that the target bypass circuit has stably carried current, and the current in the main switching circuit begins to decrease and stabilizes at a threshold, the current transfer is considered successful. A tripping command is sent to the suppression circuit contactor to disconnect it from the circuit. The operation interlock of the vacuum contactor of the main switching circuit is released.
[0023] In conjunction with the first aspect, in the sixth embodiment of the first aspect of this application, the real-time monitoring of the grid-connected voltage, and adjusting the closing rate of the main switching circuit when a sudden rise or fall in voltage occurs, includes:
[0024] The voltage values at the grid connection point are sampled. Within each analysis time window, linear regression analysis is performed on the collected voltage sequence to calculate its instantaneous slope and acceleration. A sudden drop trend indicates that the instantaneous slope is negative and the absolute value exceeds the decreasing trend threshold; a sudden rise trend indicates that the instantaneous slope is positive and the absolute value exceeds the rising trend threshold; a steady state indicates that the absolute value of the instantaneous slope is below the threshold. When a sudden drop trend is identified, the control objective is to slow down the rate of voltage drop and increase the voltage; when a sudden rise trend is identified, the control objective is to slow down the rate of voltage rise and suppress the voltage.
[0025] In conjunction with the first aspect, in the seventh embodiment of the first aspect of this application, the step of controlling the main on / off circuit to trip and disconnect the main connection between the collector line and the inverter after the shutdown command is triggered, while the bypass circuit remains operational, includes:
[0026] A tripping command is sent to the control unit of the main switching circuit to initiate the main connection disconnection process. The bypass circuit maintains its original operating state to ensure continuous current carrying. During the tripping process, the operation status of the main switching circuit is monitored in real time to confirm that its contacts gradually separate and the main connection between the collector line and the inverter is gradually disconnected. When the main switching circuit sends back a tripping signal and the current in the main switching circuit is detected to have dropped to zero, the main connection is determined to have been successfully disconnected. Simultaneously, it is confirmed that the bypass circuit is still operating normally, and the current and voltage parameters are stable and not affected by the tripping of the main switching circuit. The operation is then completed and the status is recorded.
[0027] In conjunction with the first aspect, in the eighth embodiment of the first aspect of this application, after the main switching circuit is tripped, the current and voltage status of the bypass circuit are detected, and after confirming that there is no abnormal overcurrent or overvoltage, the bypass circuit is disconnected, including:
[0028] After the main circuit trips, first confirm that the main circuit current has dropped to zero and remains stable, then start the bypass circuit status detection process; monitor the current and voltage of the bypass circuit in real time to determine if there is any abnormal overcurrent or abnormal overvoltage; if the current and voltage of the bypass circuit are within the normal range and there are no abnormal fluctuations during continuous monitoring, it is determined that the disconnection condition is met and preparation is made to execute the disconnection operation; start the bypass circuit disconnection program and control the disconnection speed according to the preset logic; after the disconnection action is completed, confirm that the bypass circuit current has dropped to zero and the voltage has returned to a stable state, and determine that the disconnection is successful.
[0029] Secondly, this application provides a photovoltaic inverter operation control system based on multi-source data, including:
[0030] Control command generation module: includes: loop construction unit and control command generation unit; wherein, in the collector line associated with the inverter, the loop construction unit constructs the main on / off loop, bypass loop and suppression loop; the control command generation unit collects multi-source data, determines whether the photovoltaic power station has the conditions for power generation, and generates control commands, including commissioning commands and shutdown commands;
[0031] The commissioning control module includes a low-load condition detection unit and an LSTM closing control unit. When the commissioning command is triggered, the low-load condition detection unit uses detection equipment to confirm whether the current on the grid side and the inverter side is in a low-load state. If the current exceeds the limit, the commissioning is delayed. The LSTM closing control unit uses an LSTM predictive model to control the closing of the vacuum contactor in the main switching circuit. It connects a suppression circuit to the circuit and uses a resistor to slow down the rate of current rise at the moment of closing.
[0032] The circuit control module includes a path optimization and switching unit and a voltage monitoring and adjustment unit. The path optimization and switching unit selects the optimal path based on the orientation relationship between the main on / off circuit and the bypass circuit and the photovoltaic power station after the main on / off circuit current stabilizes, switches to the bypass circuit, and disconnects the suppression circuit. The voltage monitoring and adjustment unit monitors the grid-connected voltage in real time and adjusts the closing rate of the main on / off circuit when a sudden rise or fall in voltage occurs.
[0033] The shutdown control module includes a shutdown control unit, a bypass circuit disconnection control unit, and a voltage fluctuation interlocking unit. When a shutdown command is triggered, the shutdown control unit controls the main circuit to trip, disconnecting the main connection between the collector line and the inverter, while the bypass circuit remains operational. After the main circuit trips, the bypass circuit disconnection control unit detects the current and voltage status of the bypass circuit. Once it confirms there is no abnormal overcurrent or overvoltage, it disconnects the bypass circuit. When the voltage fluctuation interlocking unit detects excessive voltage fluctuations, it triggers interlocking, suspending the tripping operation.
[0034] Compared with the prior art, the beneficial effects of the present invention are:
[0035] 1. This invention establishes a three-circuit architecture of main switching, bypass, and suppression, and combines multi-source data to determine power generation conditions. Before commissioning, it verifies the low-load state to avoid blind closing. It introduces an LSTM prediction model to dynamically optimize the closing rate of the vacuum contactor, and uses the suppression circuit resistor to limit current and reduce current surges.
[0036] 2. After the main switching circuit is stabilized, this invention selects the optimal path based on the orientation relationship between the main and bypass circuits and the photovoltaic array, combined with the overlap of the main power generation array, to ensure that the path matches the core area of power transmission, reduce transmission loss, and improve the operating efficiency of the power station.
[0037] 3. This invention monitors the grid-connected voltage in real time and captures sudden rises and falls, dynamically adjusts the closing rate of the main on / off circuit, and achieves proactive prevention and control through trend prediction and real-time adjustment to avoid fluctuations spreading to the power grid; when the voltage exceeds the standard during a shutdown, it triggers immediate blocking and resumes operation after stabilization, ensuring grid-friendliness. Attached Figure Description
[0038] Figure 1 This is a schematic diagram of the steps of a photovoltaic inverter operation control method based on multi-source data according to the present invention;
[0039] Figure 2 This is a system structure diagram of a photovoltaic inverter operation control system based on multi-source data according to the present invention. Detailed Implementation
[0040] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0041] Example: Figures 1-2 As shown, the present invention provides a technical solution:
[0042] like Figure 1 As shown, this application provides a photovoltaic inverter operation control method based on multi-source data, including the following steps:
[0043] Step S100: In the collector lines associated with the inverter, build the main on / off circuit, bypass circuit and suppression circuit; collect multi-source data, determine whether the photovoltaic power station has the conditions for power generation, and generate control commands, including commissioning commands and shutdown commands;
[0044] Specifically, the main switching circuit is responsible for the main control switching of the collector line during commissioning and shutdown, the bypass circuit is responsible for the current diversion after the main switching circuit stabilizes, and the suppression circuit is responsible for suppressing inrush current and voltage fluctuations during commissioning.
[0045] The main switching circuit is connected in series between the inverter output and the collector bus to form the main current path. A disconnect switch, a vacuum contactor, and a current transformer are connected in series within the circuit. Voltage monitoring points are connected in parallel across the circuit, and voltage transformers are connected to these points to monitor the voltage across the main switching circuit in real time. A bypass circuit is connected in parallel across the vacuum contactor of the main switching circuit, forming a parallel structure between the main switching circuit and the bypass circuit. A bypass disconnect switch, a bypass contactor, and a current transformer are configured within this bypass circuit, which is grouped according to the photovoltaic array orientation. A suppression circuit is connected in series between the vacuum contactor of the main switching circuit and the collector bus to form a series path. A current-limiting resistor and a suppression circuit contactor are configured within this circuit. Voltage monitoring points are connected in parallel across the suppression circuit to assist in determining the voltage division when the resistor is connected.
[0046] Furthermore, the multi-source data includes photovoltaic module operating condition data, grid connection status data, environmental interference data, and equipment health status data. Based on the photovoltaic power plant design standards and equipment parameters, an effective threshold is preset for each type of multi-source data. The multi-source data is collected, pre-processed, and compared with the effective threshold. When the module has power generation capacity, the grid has acceptance capacity, and the equipment has operational safety, an operation command is generated; otherwise, an operation shutdown command is generated.
[0047] In one specific embodiment, this embodiment is based on a 300kW photovoltaic inverter and a 10kV collector line scenario, and the experiment time is 9:10 am on a sunny day.
[0048] The main switching circuit is connected in series between the output of the 300kW inverter and the 10kV collector bus, equipped with a GN30-12 / 200 disconnecting switch (rated current 200A), a JCZ5-12D / 200 vacuum contactor (rated current 200A, opening and closing time 45ms), and an LMZJ1-0.5 current transformer (ratio 200 / 5A) connected in series. A JDZJ-10 voltage transformer is connected in parallel across both ends to monitor the voltage. The bypass circuit is grouped into east and west zones (two small arrays corresponding to the 300kW capacity). Each group is connected in parallel across the vacuum contactor, equipped with a GW9-12 / 200 disconnecting switch and a CJX2-200 contactor. The suppression circuit is connected in series between the vacuum contactor and the bus, including a 10Ω / 5kW current-limiting resistor and an LC1D200 contactor, with voltage monitoring points connected in parallel across both ends.
[0049] Preset effective thresholds: Single-module open-circuit voltage 32-40V, total array power ≥50kW, grid-connected voltage 9.5-10.5kV, frequency 49.5-50.5Hz, illuminance ≥200W / m², inverter insulation resistance ≥100MΩ. Real-time data acquisition: Single-module open-circuit voltage 36V, total array power 280kW, grid-connected voltage 10.0kV, frequency 50.1Hz, harmonic distortion rate 2.2%, illuminance 800W / m², ambient temperature 25℃, inverter insulation resistance 500MΩ, contactor coil temperature 40℃. After data preprocessing, thresholds are compared: If the modules, grid, and equipment all meet the standards, a commissioning command is generated.
[0050] Step S200: After the commissioning command is triggered, the detection equipment is used to confirm whether the current on the grid side and the inverter side is in a low load state. If the current exceeds the standard, the commissioning is delayed. The LSTM prediction model is used to control the vacuum contactor of the main switching circuit to close, and the suppression circuit is connected to the circuit to slow down the current rise rate at the moment of closing by using a resistor.
[0051] Specifically, based on the inverter's rated parameters, collector line design standards, and inrush current suppression targets, a quantitative threshold for low-load conditions is set; the detection equipment on the grid side and inverter side is pre-calibrated; when the commissioning command is triggered, the current on the grid side and inverter side is collected and pre-processed; when the grid side current is not greater than the low-load threshold and the inverter side current is not greater than the low-load threshold, it is determined that the system is in a low-load state; otherwise, the current is determined to be excessive; when the current is excessive, the closing action of the main switching circuit is frozen, and an initial delay time is set; after the delay time expires, it is re-evaluated whether the low-load requirements are met; if the current is still excessive, the verification process is repeated until the low-current operating condition requirements are met.
[0052] Furthermore, time-series data of past closing processes of photovoltaic power plants are collected to construct a training sample library. Input time-series features include grid-side voltage sequences, inverter-side initial current sequences, irradiance sequences, module temperature sequences, and real-time temperature sequences of suppression resistors. Output time-series labels, specifically the current rise sequence after closing. The data is cleaned and normalized, and divided into training and validation sets. An LSTM prediction model is constructed with the goal of minimizing the mean square error between the predicted current sequence and the actual sequence. The Adam optimizer is used for training to ensure that the prediction error is within the allowable range. The collected current time-series features are input into the LSTM model, and the prediction results, specifically the current rise curve, are output. Based on the curve, the predicted peak value, predicted rise slope, and predicted core saturation time are extracted to generate control parameters for the closing and suppression circuits. Based on the control parameters, a stepped closing command is sent to the vacuum contactor, and an early closing command is sent to the suppression circuit contactor.
[0053] In one specific embodiment, this embodiment is based on a 300kW photovoltaic inverter and a 10kV collector line scenario, and the commissioning command is triggered at 9:15 on a sunny day (receiving the S100 commissioning command).
[0054] During the low-load status confirmation phase, based on the rated current of the 300kW inverter (17.32A, 10kV system), the low-load threshold is set to 10% of the rated value, i.e., 1.75A. The LMZJ1-0.5 type current transformers (200 / 5A ratio) on both the grid side and the inverter side are pre-calibrated, achieving an accuracy of 0.5 class after calibration. After the commissioning command is triggered, the grid-side current is collected at 1.2A and the inverter-side current at 0.8A. After filtering and preprocessing, both are below 1.75A, indicating that the low-load status meets the standard. Counterexample: If the grid-side current is 2.0A (exceeding the standard), the circuit breaker is frozen and a 10-second delay is set for retesting. Once the second collected value of 1.3A meets the standard, the process continues.
[0055] In the LSTM control phase, the training sample library consists of 800 past closing timing data (50ms step size). Input features include grid voltage fluctuations of 10.0-10.05kV, inverter initial current of 0.8A, solar irradiance of 800W / m², module temperature of 25℃, and suppression resistor (10Ω / 5kW) temperature of 23℃. The model is a 2-layer LSTM (64+32 nodes), trained with an Adam optimizer (learning rate 0.001), and a validation set MSE of 0.006A². After the current features are input, the model predicts a peak current of 16A, a rising slope of 0.18A / ms, and a saturation time of 30ms after closing. The generated control parameters are: vacuum contactor stepped closing rate of 0.12 seconds, and suppression circuit contactor closing 5ms in advance. After execution, the measured peak current is 16.3A, and the slope is 0.19A / ms, meeting the error requirements.
[0056] Step S300: After the main switching circuit current stabilizes, select the optimal path based on the orientation relationship between the main switching circuit and the bypass circuit and the photovoltaic power station, switch to bypass circuit operation, and disconnect the suppression circuit; monitor the grid voltage in real time, and adjust the closing rate of the main switching circuit when a sudden rise or fall in voltage occurs.
[0057] Specifically, the pre-stored power plant topology data is called. This data defines the physical connection relationship and electrical path between all bypass circuits and each group of photovoltaic arrays. The data attributes include circuit identifier, associated array group, azimuth coordinates and current status. The associated array group is a subset of photovoltaic arrays connected to the circuit. The azimuth coordinates are the relative electrical distance between the circuit access point and the array group. The current status is the status of the disconnecting switch and contactor of the bypass circuit.
[0058] The selection principle is to choose the bypass circuit that is closest to the current main power generation array in terms of location. Real-time power generation data of each photovoltaic array group is obtained, and one or more array groups contributing the main power generation are identified. The identified main power generation array groups are matched with the array groups associated with each bypass circuit. The path priority score of each bypass circuit is calculated, with the scoring logic being: if the array group associated with a bypass circuit has the greatest overlap with the current main power generation array group, its score is the highest. The bypass circuit with the highest path priority score is selected as the target bypass circuit for this switchover. When multiple circuits with the same score appear, the one with the shortest electrical path is selected. An operation interlock signal is sent to the vacuum contactor of the main switching circuit to close the bypass isolating switch of the target bypass circuit. Subsequently, the bypass contactor of that circuit is closed. When it is confirmed that the target bypass circuit has stably carried current, and the current in the main switching circuit begins to decrease and stabilizes at a threshold, the current transfer is considered successful. A tripping command is sent to the suppression circuit contactor to disconnect it from the circuit. The operation interlock of the vacuum contactor of the main switching circuit is released.
[0059] Furthermore, the voltage values at the grid connection point are sampled, and linear regression analysis is performed on the collected voltage sequence within each analysis time window to calculate its instantaneous slope and acceleration. A sudden drop trend indicates that the instantaneous slope is negative and the absolute value exceeds the drop trend threshold; a sudden rise trend indicates that the instantaneous slope is positive and the absolute value exceeds the rise trend threshold; a steady state indicates that the absolute value of the instantaneous slope is below the threshold. When a sudden drop trend is identified, the control objective is to slow down the rate of voltage drop and increase the voltage; when a sudden rise trend is identified, the control objective is to slow down the rate of voltage rise and suppress the voltage.
[0060] In one specific embodiment, this embodiment is based on a 300kW photovoltaic inverter and a 10kV collector line scenario. The stabilization time after the main switching circuit is closed is 9:20, at which time the main switching circuit current is 17A (fluctuation ≤2%).
[0061] Pre-stored power station topology data: The bypass circuits are divided into east and west groups. The east bypass is associated with the east zone array at an electrical distance of 100m, and the west bypass is associated with the west zone array at a distance of 130m. Both groups are in standby mode with "disconnector open and contactor open". Real-time array power acquisition: East zone 160kW, West zone 120kW, identifying the east zone as the main power generation area (accounting for 53.3%). Score calculation: The east bypass is associated with the main array (overlap 100%, score 1), the west bypass has no overlap (score 0), and the east bypass is selected as the target. Execution of switching: A blocking signal is sent to the main circuit disconnection circuit, the east bypass disconnector is closed (feedback 18ms), and then the contactor is closed; after 300ms, the east bypass current is monitored to rise to 16.5A, and the main circuit disconnection circuit current drops to 8A (stable ≤10A), indicating successful transfer; a tripping command is sent to the suppression circuit (tripping time 40ms), releasing the main circuit disconnection circuit blocking.
[0062] A voltage surge threshold was set: instantaneous slope > 0.003 kV / ms, with an analysis window of 50 ms. During the switching process, the grid-connected voltage was collected in real time: rising from 10.0 kV to 10.2 kV. Linear regression calculation showed an instantaneous slope of 0.004 kV / ms (exceeding the threshold, indicating a surge). The closing rate of the main switching circuit was immediately adjusted (from 0.12 seconds to 0.18 seconds). After 100 ms, the voltage stabilized at 10.1 kV, and the slope decreased to 0.002 kV / ms (returning to stability).
[0063] Step S400: When the shutdown command is triggered, the main switching circuit is opened, disconnecting the main connection between the collector line and the inverter, while the bypass circuit remains in operation. After the main switching circuit is opened, the current and voltage status of the bypass circuit are detected. After confirming that there is no abnormal overcurrent or overvoltage, the bypass circuit is disconnected. When voltage fluctuations exceed the limit, the interlocking is triggered, and the opening operation is suspended.
[0064] Specifically, a tripping command is sent to the control unit of the main switching circuit to initiate the main connection disconnection process. The bypass circuit maintains its original operating state to ensure continuous current carrying. During the tripping process, the operation status of the main switching circuit is monitored in real time to confirm that its contacts gradually separate and the main connection between the collector line and the inverter is gradually disconnected. When the main switching circuit sends back a tripping signal and the current in the main switching circuit is detected to have dropped to zero, the main connection is determined to have been successfully disconnected. Simultaneously, it is confirmed that the bypass circuit is still operating normally, and the current and voltage parameters are stable and not affected by the tripping of the main switching circuit. The operation is then completed and the status is recorded.
[0065] Furthermore, after the main circuit trips, it is first confirmed that the main circuit current has dropped to zero and remained stable. Then, the bypass circuit status detection process is initiated. The current and voltage of the bypass circuit are monitored in real time to determine whether there is any abnormal overcurrent or abnormal overvoltage. If the current and voltage of the bypass circuit are within the normal range and there are no abnormal fluctuations during continuous monitoring, it is determined that the disconnection condition is met and the disconnection operation is prepared. The bypass circuit disconnection program is initiated, and the disconnection speed is controlled according to the preset logic. After the disconnection action is completed, it is confirmed that the bypass circuit current has dropped to zero and the voltage has returned to a stable state, and the disconnection is determined to be successful.
[0066] In one specific embodiment, this embodiment is based on a 300kW photovoltaic inverter and a 10kV collector line scenario, and the shutdown command is triggered at 17:40 in the evening (when the light intensity drops to 190W / m²).
[0067] A tripping command is sent to the main on / off circuit control unit, and the JCZ5-12D / 200 vacuum contactor initiates tripping (action time 45ms). During the tripping process, the main on / off circuit current is monitored: it gradually decreases from 16.5A, dropping to 8A at 25ms, and then to 0 at 45ms. This current remains at zero for 50ms, confirming that the main connection is disconnected. Simultaneously, the east bypass circuit is monitored: the current rises from 0 to 16.2A (fluctuation ≤3%), and the voltage is 10.0kV (stabilizing in the 9.5-10.5kV range), confirming normal operation.
[0068] After the main circuit trips, the bypass circuit is activated for detection: real-time current is 16.2A (normal range 0-18A) and voltage is 10.0kV (no overvoltage / undervoltage), with no fluctuations for 200ms, meeting the tripping conditions. Tripping is initiated according to the preset logic (0.15-second tripping speed). During this process, the current smoothly decreases from 16.2A to 0 (taking 150ms), and the voltage eventually stabilizes at 10.1kV, indicating successful tripping.
[0069] During simulated circuit breaker tripping, the grid disturbance occurs: the voltage suddenly rises to 10.6kV (exceeding the 10.5kV threshold), immediately triggering a lockout and suspending the bypass disconnection; after 1 second, the voltage drops back to 10.2kV and stabilizes, the lockout is released, and the disconnection resumes from the paused step, finally completing the operation.
[0070] like Figure 2 As shown, this application provides a photovoltaic inverter operation control system based on multi-source data, including:
[0071] Control command generation module: includes: loop construction unit and control command generation unit; wherein, in the collector line associated with the inverter, the loop construction unit constructs the main on / off loop, bypass loop and suppression loop; the control command generation unit collects multi-source data, determines whether the photovoltaic power station has the conditions for power generation, and generates control commands, including commissioning commands and shutdown commands;
[0072] The commissioning control module includes a low-load condition detection unit and an LSTM closing control unit. When the commissioning command is triggered, the low-load condition detection unit uses detection equipment to confirm whether the current on the grid side and the inverter side is in a low-load state. If the current exceeds the limit, the commissioning is delayed. The LSTM closing control unit uses an LSTM predictive model to control the closing of the vacuum contactor in the main switching circuit. It connects a suppression circuit to the circuit and uses a resistor to slow down the rate of current rise at the moment of closing.
[0073] The circuit control module includes a path optimization and switching unit and a voltage monitoring and adjustment unit. The path optimization and switching unit selects the optimal path based on the orientation relationship between the main on / off circuit and the bypass circuit and the photovoltaic power station after the main on / off circuit current stabilizes, switches to the bypass circuit, and disconnects the suppression circuit. The voltage monitoring and adjustment unit monitors the grid-connected voltage in real time and adjusts the closing rate of the main on / off circuit when a sudden rise or fall in voltage occurs.
[0074] The shutdown control module includes a shutdown control unit, a bypass circuit disconnection control unit, and a voltage fluctuation interlocking unit. When a shutdown command is triggered, the shutdown control unit controls the main circuit to trip, disconnecting the main connection between the collector line and the inverter, while the bypass circuit remains operational. After the main circuit trips, the bypass circuit disconnection control unit detects the current and voltage status of the bypass circuit. Once it confirms there is no abnormal overcurrent or overvoltage, it disconnects the bypass circuit. When the voltage fluctuation interlocking unit detects excessive voltage fluctuations, it triggers interlocking, suspending the tripping operation.
[0075] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.
Claims
1. A photovoltaic inverter operation control method based on multi-source data, characterized in that, Includes the following steps: In the collector lines associated with the inverter, main switching circuit, bypass circuit and suppression circuit are built; multi-source data are collected to determine whether the photovoltaic power station has the conditions for power generation and generate control commands, including commissioning commands and shutdown commands; When the commissioning command is triggered, the detection equipment confirms whether the current on the grid side and the inverter side is in a low load state. If the current exceeds the standard, the commissioning is delayed. The LSTM prediction model is used to control the vacuum contactor of the main switching circuit to close, and the suppression circuit is connected to the circuit to slow down the current rise rate at the moment of closing by using a resistor. The method of using an LSTM predictive model to control the closing of the vacuum contactor in the main switching circuit includes connecting a suppression circuit to the circuit and using a resistor to slow down the rate of current rise at the moment of closing, including: Time-series data of past closing processes in photovoltaic power plants are collected to construct a training sample library. Input time-series features include grid-side voltage sequences, inverter-side initial current sequences, irradiance sequences, module temperature sequences, and real-time temperature sequences of suppression resistors. Output time-series labels, specifically the current rise sequence after closing. The data is cleaned and normalized, and divided into training and validation sets. An LSTM prediction model is constructed with the goal of minimizing the mean square error between the predicted current sequence and the actual sequence. The Adam optimizer is used for training to ensure that the prediction error is within the allowable range. The collected current time-series features are input into the LSTM model, and the prediction results, specifically the current rise curve, are output. Based on the curve, the predicted peak value, predicted rise slope, and predicted core saturation time are extracted to generate control parameters for the closing and suppression circuits. Based on the control parameters, a stepped closing command is sent to the vacuum contactor, and an early closing command is sent to the suppression circuit contactor. After the main switching circuit current stabilizes, the optimal path is selected based on the orientation relationship between the main switching circuit and the bypass circuit and the photovoltaic power station, and the system is switched to the bypass circuit to operate, while the suppression circuit is disconnected. The grid-connected voltage is monitored in real time, and the closing rate of the main switching circuit is adjusted when a sudden rise or fall in voltage occurs. The orientation relationship includes the relative electrical distance between the circuit and the photovoltaic array group, and the degree of overlap between the array group associated with the circuit and the main power generation array group. When the shutdown command is triggered, the main on / off circuit is opened, disconnecting the main connection between the collector line and the inverter, while the bypass circuit remains operational. After the main on / off circuit is opened, the current and voltage status of the bypass circuit are detected. Once it is confirmed that there is no abnormal overcurrent or overvoltage, the bypass circuit is disconnected. When voltage fluctuations exceed the limit, the interlocking is triggered, and the opening operation is suspended.
2. The photovoltaic inverter operation control method based on multi-source data according to claim 1, characterized in that, The construction of a main switching circuit, a bypass circuit, and a suppression circuit in the collector lines associated with the inverter includes: The main switching circuit is responsible for the main control switching of the collector line during commissioning and shutdown. The bypass circuit is responsible for the current diversion after the main switching circuit stabilizes. The suppression circuit is responsible for suppressing inrush current and voltage fluctuations during commissioning. The main switching circuit is connected in series between the inverter output and the collector bus to form the main current path. A disconnect switch, a vacuum contactor, and a current transformer are connected in series within the circuit. Voltage monitoring points are connected in parallel across the circuit, and voltage transformers are connected to these points to monitor the voltage across the main switching circuit in real time. A bypass circuit is connected in parallel across the vacuum contactor of the main switching circuit, forming a parallel structure between the main switching circuit and the bypass circuit. A bypass disconnect switch, a bypass contactor, and a current transformer are configured within this bypass circuit, which is grouped according to the photovoltaic array orientation. A suppression circuit is connected in series between the vacuum contactor of the main switching circuit and the collector bus to form a series path. A current-limiting resistor and a suppression circuit contactor are configured within this circuit. Voltage monitoring points are connected in parallel across the suppression circuit to assist in determining the voltage division when the resistor is connected.
3. The photovoltaic inverter operation control method based on multi-source data according to claim 1, characterized in that, The process involves collecting multi-source data to determine whether the photovoltaic power station meets the conditions for power generation, and generating control commands, including commissioning commands and shutdown commands. Multi-source data includes photovoltaic module operating condition data, grid connection status data, environmental interference data, and equipment health status data. Based on photovoltaic power plant design standards and equipment parameters, effective thresholds for each type of multi-source data are preset. Multi-source data is collected, pre-processed, and compared with the effective thresholds. When the module has power generation capacity, the grid has acceptance capacity, and the equipment has operational safety, a commissioning command is generated; otherwise, a shutdown command is generated.
4. The photovoltaic inverter operation control method based on multi-source data according to claim 1, characterized in that, When the commissioning command is triggered, the detection equipment confirms whether the current on the grid side and the inverter side is in a low-load state. If the current exceeds the standard, the commissioning is delayed, including: Based on the inverter's rated parameters, collector line design standards, and inrush current suppression targets, a quantification threshold for low-load conditions is set. Pre-calibration is performed on the detection equipment on both the grid and inverter sides. Upon triggering the commissioning command, current acquisition on both the grid and inverter sides is initiated, and data preprocessing is performed. If both the grid-side and inverter-side currents are not greater than the low-load threshold, the system is determined to be in a low-load state; otherwise, the current is deemed to be exceeding the limit. If the current exceeds the limit, the main switching circuit closing action is frozen, and an initial delay time is set. After the delay time expires, the system re-evaluates whether the low-load requirements are met. If the current still exceeds the limit, the verification process is repeated until the low-current operating condition requirements are met.
5. The photovoltaic inverter operation control method based on multi-source data according to claim 1, characterized in that, After the current in the main switching circuit stabilizes, the optimal path is selected based on the location relationship between the main switching circuit and the bypass circuit and the photovoltaic power station, and the system switches to bypass circuit operation, disconnecting the suppression circuit, including: The system calls up pre-stored power plant topology data, which defines the physical connection relationships and electrical paths between all bypass circuits and each group of photovoltaic arrays. The data attributes include circuit identifier, associated array group, azimuth coordinates, and current status. The associated array group is a subset of photovoltaic arrays connected to the bypass circuit. The azimuth coordinates are the relative electrical distance between the circuit access point and the array group. The current status is the status of the disconnector and contactor of the bypass circuit. The selection principle is to choose the bypass circuit that is closest to the current main power generation array in terms of location. Real-time power generation data of each photovoltaic array group is obtained, and one or more array groups contributing to power generation are identified. The identified main power generation array groups are matched with the array groups associated with each bypass circuit. The path priority score of each bypass circuit is calculated, with the scoring logic being: if the array group associated with a bypass circuit has the greatest overlap with the current main power generation array group, its score is the highest. The bypass circuit with the highest path priority score is selected as the target bypass circuit for this switchover. When multiple circuits with the same score appear, the one with the shortest electrical path is selected. An operation interlock signal is sent to the vacuum contactor of the main switching circuit to close the bypass isolating switch of the target bypass circuit. Subsequently, the bypass contactor of the target bypass circuit is closed. When it is confirmed that the target bypass circuit has stably carried current, and the current in the main switching circuit begins to decrease and stabilizes at a threshold, the current transfer is considered successful. A tripping command is sent to the suppression circuit contactor to disconnect it from the circuit. The operation interlock of the vacuum contactor of the main switching circuit is released.
6. The photovoltaic inverter operation control method based on multi-source data according to claim 1, characterized in that, The real-time monitoring of grid-connected voltage, when a sudden rise or fall in voltage occurs, adjusts the closing rate of the main switching circuit, including: The voltage values at the grid connection point are sampled. Within each analysis time window, linear regression analysis is performed on the collected voltage sequence to calculate its instantaneous slope and acceleration. A sudden drop trend indicates that the instantaneous slope is negative and the absolute value exceeds the decreasing trend threshold; a sudden rise trend indicates that the instantaneous slope is positive and the absolute value exceeds the rising trend threshold; a steady state indicates that the absolute value of the instantaneous slope is below the threshold. When a sudden drop trend is identified, the control objective is to slow down the rate of voltage drop and increase the voltage; when a sudden rise trend is identified, the control objective is to slow down the rate of voltage rise and suppress the voltage.
7. The photovoltaic inverter operation control method based on multi-source data according to claim 1, characterized in that, When the shutdown command is triggered, the main on / off circuit is tripped, disconnecting the main connection between the collector line and the inverter, while the bypass circuit remains operational, including: A tripping command is sent to the control unit of the main switching circuit to initiate the main connection disconnection process. The bypass circuit maintains its original operating state to ensure continuous current carrying. During the tripping process, the operation status of the main switching circuit is monitored in real time to confirm that its contacts gradually separate and the main connection between the collector line and the inverter is gradually disconnected. When the main switching circuit sends back a tripping signal and the current in the main switching circuit is detected to have dropped to zero, the main connection is determined to have been successfully disconnected. Simultaneously, it is confirmed that the bypass circuit is still operating normally, and the current and voltage parameters are stable and not affected by the tripping of the main switching circuit. The operation is then completed and the status is recorded.
8. The photovoltaic inverter operation control method based on multi-source data according to claim 1, characterized in that, After the main switching circuit is tripped, the current and voltage status of the bypass circuit are detected. After confirming that there is no abnormal overcurrent or overvoltage, the bypass circuit is disconnected, including: After the main circuit trips, first confirm that the main circuit current has dropped to zero and remains stable, then start the bypass circuit status detection process; monitor the current and voltage of the bypass circuit in real time to determine if there is any abnormal overcurrent or abnormal overvoltage; if the current and voltage of the bypass circuit are within the normal range and there are no abnormal fluctuations during continuous monitoring, it is determined that the disconnection condition is met and preparation is made to execute the disconnection operation; start the bypass circuit disconnection program and control the disconnection speed according to the preset logic; after the disconnection action is completed, confirm that the bypass circuit current has dropped to zero and the voltage has returned to a stable state, and determine that the disconnection is successful.
9. A photovoltaic inverter operation control system based on multi-source data, using the photovoltaic inverter operation control method based on multi-source data according to any one of claims 1-8, characterized in that, include: Control command generation module: includes: loop construction unit and control command generation unit; wherein, in the collector line associated with the inverter, the loop construction unit constructs the main on / off loop, bypass loop and suppression loop; the control command generation unit collects multi-source data, determines whether the photovoltaic power station has the conditions for power generation, and generates control commands, including commissioning commands and shutdown commands; The commissioning control module includes a low-load condition detection unit and an LSTM closing control unit. When the commissioning command is triggered, the low-load condition detection unit uses detection equipment to confirm whether the current on the grid side and the inverter side is in a low-load state. If the current exceeds the limit, the commissioning is delayed. The LSTM closing control unit uses an LSTM predictive model to control the closing of the vacuum contactor in the main switching circuit. It connects a suppression circuit to the circuit and uses a resistor to slow down the rate of current rise at the moment of closing. The loop control module includes a path optimization and switching unit and a voltage monitoring and adjustment unit. The path optimization and switching unit, after the main switching loop current stabilizes, selects the optimal path based on the orientation relationship between the main switching loop and the bypass loop and the photovoltaic power station, switches to the bypass loop operation, and disconnects the suppression loop. The voltage monitoring and adjustment unit monitors the grid-connected voltage in real time, and adjusts the closing rate of the main switching loop when a sudden rise or fall in voltage occurs. The orientation relationship includes the relative electrical distance between the loop and the photovoltaic array group, and the degree of overlap between the array group associated with the loop and the main power generation array group. The shutdown control module includes a shutdown control unit, a bypass circuit disconnection control unit, and a voltage fluctuation interlocking unit. When a shutdown command is triggered, the shutdown control unit controls the main circuit to trip, disconnecting the main connection between the collector line and the inverter, while the bypass circuit remains operational. After the main circuit trips, the bypass circuit disconnection control unit detects the current and voltage status of the bypass circuit. Once it confirms there is no abnormal overcurrent or overvoltage, it disconnects the bypass circuit. When the voltage fluctuation interlocking unit detects excessive voltage fluctuations, it triggers interlocking, suspending the tripping operation.