An active power output scheduling system and method for a photovoltaic power station

Abstract

This invention relates to the field of photovoltaic power plant technology, and discloses an active power output scheduling system and method for photovoltaic power plants. The system comprises a module that acquires real-time meteorological monitoring data and electrical operation data of the photovoltaic power plant to obtain a multi-source time-series dataset; a power prediction module that generates an ultra-short-term power prediction sequence and a short-term power prediction sequence based on the multi-source time-series dataset; a deviation calculation module that calculates the active power deviation based on the active power scheduling limit and the ultra-short-term power prediction sequence; a power setting module that determines the active power setpoint for each inverter unit in response to the active power deviation exceeding the dynamic adjustment dead zone; a power output module that converts the active power setpoint into a modulation command and sends it to the corresponding inverter to monitor the actual output power; and a power correction module that corrects the prediction deviation coefficient for the next control cycle, optimizing the active power output scheduling and improving the power prediction accuracy, scheduling response speed, and grid active support capability of the photovoltaic power plant.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic power plant technology, and more specifically, to an active power output scheduling system and method for a photovoltaic power plant. Background Technology

[0002] Active power output dispatching of photovoltaic power plants refers to adjusting the output power of photovoltaic inverters in real time according to grid dispatch instructions and meteorological conditions to match the power output of the power plant with the grid demand. It is a key technology to ensure the grid connection safety of photovoltaic power plants and the stable operation of the grid.

[0003] Existing photovoltaic (PV) dispatching methods primarily employ centralized power control, where the power plant monitoring system uniformly calculates the power allocation for each inverter and issues commands through simple proportional allocation or polling methods. However, existing technologies struggle to adapt to the complex electrical topologies and line loss variations of large-scale PV power plants, lack multi-timescale power prediction capabilities, and cannot optimize dispatching based on the actual efficiency characteristics of inverters. Furthermore, existing methods mostly employ open-loop control, lacking execution feedback and prediction correction mechanisms, resulting in low dispatching accuracy and lag response, making it difficult to meet the grid's requirements for rapid power regulation and proactive support capabilities for PV power plants. Summary of the Invention

[0004] To address the aforementioned technical issues, this application aims to provide an active power output scheduling system and method for photovoltaic power plants. By integrating meteorological and electrical multi-source data to construct multi-timescale power prediction, and based on hierarchical collaborative scheduling to achieve refined power allocation at the cluster and inverter levels, the system achieves adaptive optimization of the prediction model through a closed-loop feedback mechanism, thereby improving the power prediction accuracy, scheduling response speed, and grid active support capabilities of photovoltaic power plants.

[0005] To achieve the above objectives, the present invention provides an active power output dispatching system for a photovoltaic power plant, comprising: The collection builds a module to acquire real-time meteorological monitoring data and electrical operation data of each photovoltaic array in the photovoltaic power station, and obtain a multi-source time series dataset; The power prediction module is used to perform a multi-timescale power prediction process based on the multi-source time series dataset to generate ultra-short-term power prediction sequences and short-term power prediction sequences. The deviation calculation module is used to calculate the active power deviation of the current control cycle based on the active power dispatch limit issued by the power grid dispatching agency and the ultra-short-term power prediction sequence. The power setting module is used to trigger the hierarchical collaborative scheduling process and determine the active power setting value of each inverter unit in response to the active power deviation exceeding the dynamic adjustment dead zone. The power output module is used to convert the active power setpoint into a modulation command and send it to the corresponding inverter, monitor the actual output power after execution, and generate closed-loop feedback information. The power correction module is used to correct the prediction deviation coefficient of the next control cycle based on the closed-loop feedback information, thereby optimizing the scheduling accuracy of subsequent active power output.

[0006] Furthermore, the collection building module is used for: The total irradiance and scattered irradiance of the array tilt surface are collected by an irradiance sensor, and the back panel temperature of the photovoltaic panel and the ambient temperature are collected based on a temperature sensor. Collect the DC side input voltage and input current, AC side three-phase output voltage and output current of each inverter, and calculate the instantaneous output power; Record the active power, reactive power, grid connection point voltage, and grid connection point current on the high-voltage side of the transformer substation as grid connection status parameters; The real-time meteorological monitoring data and electrical operation data are time-stamped according to a unified sampling period, and the time series are aligned to obtain the multi-source time series dataset.

[0007] Furthermore, the power prediction module is used for: Feature extraction is performed on the multi-source time series dataset to obtain trend components and volatility components; The trend component and the fluctuation component are sliced ​​according to the time window. The trend component value corresponding to each prediction time is input into the prediction branch based on the physical model. Combined with the photovoltaic module temperature correction coefficient and line loss, the first power prediction value is generated. The volatility component values ​​corresponding to each prediction time are input into the data-driven prediction branch, and a second power prediction value is generated using historical similar day data. The first power prediction value and the second power prediction value are weighted and fused to obtain the ultra-short-term power prediction sequence within a preset time period and the short-term power prediction sequence within the next few hours.

[0008] Furthermore, the deviation calculation module is used for: Analyze multiple active power dispatch limits issued by the power grid dispatching agency and extract the target output value of the planned processing curve for the corresponding period in the next control cycle; Extract the predicted output value with the same control period from the ultra-short-term power prediction sequence; Calculate the difference between the predicted output value and the target output value, and use it as the basic deviation. Collect real-time evaluation deviations of grid connection points and power deviations of tie lines, and calculate power compensation based on grid frequency regulation requirements; The active power deviation is obtained by vector addition of the basic deviation and the power compensation.

[0009] Furthermore, the power setting module is used for: Based on the electrical topology of the photovoltaic power station, the inverters are divided into multiple clusters, with each cluster corresponding to a collector line; Calculate the current available power generation margin of each cluster and the line loss coefficient to the grid connection point, and construct a cluster adjustment priority sequence; According to the cluster adjustment priority sequence, the active power deviation is allocated to each cluster in sequence until the deviation is fully allocated or the cluster margin is exhausted. For each cluster that has obtained the deviation allocation, the active power setpoint of a single inverter is calculated based on the current operating status and efficiency characteristic curve of each inverter in the cluster.

[0010] Furthermore, the power setting module is used for: Obtain the rated capacity and current actual output power of each inverter in the cluster, and calculate the difference to obtain the remaining adjustable capacity of a single unit. The total adjustable margin of the cluster is obtained by counting the number of inverters in grid-connected operation within the cluster and summing the remaining adjustable capacity of each individual unit. The system collects the real-time current carrying capacity and current transmission power of the corresponding power collection lines of the data acquisition cluster, and calculates the remaining transmission capacity of the lines. Compare the total adjustable margin of the cluster with the remaining transmission capacity of the line, and take the minimum of the two as the current available power generation margin.

[0011] Furthermore, the power setting module is used for: Obtain the maximum power point tracking efficiency and conversion efficiency of each inverter in the cluster under the current irradiance and temperature conditions; The efficiency values ​​are normalized to eliminate dimensional differences and then used as weighting coefficients. The power deviation that the cluster needs to bear is allocated proportionally according to the proportion of each inverter's weight coefficient to the total weight of the cluster. The allocated power regulation is algebraically calculated with the corresponding inverter's current actual output power to generate the active power setpoint, which is then constrained within the inverter's allowable output range.

[0012] Furthermore, the power output module is used for: The active power setpoint is converted into a reference value for the AC side current amplitude of the inverter, and a current vector command is generated by combining the current grid voltage phase information. The current vector command is decomposed into active current components and reactive current components, and encapsulated into a standard remote control message through a communication protocol. The remote control message is transmitted to the target inverter via a fiber optic ring network or power line carrier communication channel; After receiving the message, the inverter analyzes and extracts the active current component, and generates a pulse width modulation signal through proportional-integral operation.

[0013] Furthermore, the power output module is used for: The active power setpoint is converted into a reference value for the AC side current amplitude of the inverter, and a current vector command is generated by combining the current grid voltage phase information. The current vector command is decomposed into active current components and reactive current components, and encapsulated into a standard remote control message through a communication protocol. The remote control message is transmitted to the target inverter via a fiber optic ring network or power line carrier communication channel; After receiving the message, the inverter analyzes and extracts the active current component, and generates a pulse width modulation signal through proportional-integral operation.

[0014] Furthermore, the power correction module is used for: The actual output power value of the inverter after adjustment is collected and compared with the active power set value to calculate the tracking error; The tracking error is correlated with the predicted value of the corresponding time period in the ultra-short-term power prediction sequence to identify the direction of systematic deviation of the prediction model. The prediction deviation coefficient is updated using an exponential smoothing algorithm, so that the corrected prediction value gradually converges to the actual output direction. The updated prediction deviation coefficients are stored in the parameter library for use in the power prediction process of the next control cycle, forming the closed-loop feedback information.

[0015] To achieve the above objectives, the present invention also provides a method for active power output scheduling of a photovoltaic power plant, comprising: Real-time meteorological monitoring data and electrical operation data of each photovoltaic array in the photovoltaic power station are obtained to obtain a multi-source time series dataset; Based on the multi-source time-series dataset, a multi-time-scale power prediction process is executed to generate ultra-short-term power prediction sequences and short-term power prediction sequences. Based on the active power dispatch limit issued by the power grid dispatching agency and the ultra-short-term power prediction sequence, calculate the active power deviation for the current control cycle. In response to the active power deviation exceeding the dynamic adjustment dead zone, a hierarchical collaborative scheduling process is triggered to determine the active power setpoint for each inverter unit. The active power setpoint is converted into a modulation command and sent to the corresponding inverter. The actual output power after execution is monitored, and closed-loop feedback information is generated. The prediction deviation coefficient for the next control cycle is corrected based on the closed-loop feedback information, thereby optimizing the subsequent active power output scheduling accuracy.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention discloses an active power output scheduling system and method for photovoltaic power plants. The system comprises: a construction module that acquires real-time meteorological monitoring data and electrical operation data of the photovoltaic power plant to obtain a multi-source time-series dataset; a power prediction module that generates ultra-short-term power prediction sequences and short-term power prediction sequences based on the multi-source time-series dataset; a deviation calculation module that calculates the active power deviation based on the active power scheduling limit and the ultra-short-term power prediction sequences; a power setting module that determines the active power setpoint for each inverter unit in response to the active power deviation exceeding the dynamic adjustment dead zone; a power output module that converts the active power setpoint into modulation commands and sends them to the corresponding inverters, monitoring the actual output power; and a power correction module that corrects the prediction deviation coefficient for the next control cycle, optimizing active power output scheduling and improving the power prediction accuracy, scheduling response speed, and grid active support capability of the photovoltaic power plant. Attached Figure Description

[0017] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 A schematic diagram of the active power output dispatching system of a photovoltaic power station according to an embodiment of the present invention is shown; Figure 2 A flowchart illustrating an active power output scheduling method for a photovoltaic power plant according to an embodiment of the present invention is shown. Detailed Implementation

[0018] The specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and are not intended to limit the scope of the invention.

[0019] In the description of this application, it should be understood that the terms "center", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.

[0020] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0021] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection between two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0022] The following is a description of preferred embodiments of the present invention in conjunction with the accompanying drawings.

[0023] like Figure 1 As shown, an embodiment of the present invention discloses an active power output dispatching system for a photovoltaic power plant, comprising: The collection builds a module to acquire real-time meteorological monitoring data and electrical operation data of each photovoltaic array in the photovoltaic power station, and obtain a multi-source time series dataset; The power prediction module is used to perform a multi-timescale power prediction process based on the multi-source time series dataset to generate ultra-short-term power prediction sequences and short-term power prediction sequences. The deviation calculation module is used to calculate the active power deviation of the current control cycle based on the active power dispatch limit issued by the power grid dispatching agency and the ultra-short-term power prediction sequence. The power setting module is used to trigger the hierarchical collaborative scheduling process and determine the active power setting value of each inverter unit in response to the active power deviation exceeding the dynamic adjustment dead zone. The power output module is used to convert the active power setpoint into a modulation command and send it to the corresponding inverter, monitor the actual output power after execution, and generate closed-loop feedback information. The power correction module is used to correct the prediction deviation coefficient of the next control cycle based on the closed-loop feedback information, thereby optimizing the scheduling accuracy of subsequent active power output.

[0024] In some embodiments of this application, the collection building module is used for: The total irradiance and scattered irradiance of the array tilt surface are collected by an irradiance sensor, and the back panel temperature of the photovoltaic panel and the ambient temperature are collected based on a temperature sensor. Collect the DC side input voltage and input current, AC side three-phase output voltage and output current of each inverter, and calculate the instantaneous output power; Record the active power, reactive power, grid connection point voltage, and grid connection point current on the high-voltage side of the transformer substation as grid connection status parameters; The real-time meteorological monitoring data and electrical operation data are time-stamped according to a unified sampling period, and the time series are aligned to obtain the multi-source time series dataset.

[0025] In this embodiment, the irradiance sensor uses a thermopile type or a photoelectric radiometer. The DC side input voltage and current are obtained through the inverter's own sampling circuit, and the AC side output three-phase voltage and current are obtained through an AC sampling transformer. The instantaneous output power is calculated as the sum of the products of the three-phase voltage and current instantaneous values. The active and reactive power on the high-voltage side of the transformer substation are obtained through a high-voltage side power quality monitoring device, and the grid connection point voltage and frequency are obtained through a synchronous phasor measurement unit (PMU) or energy meter at the grid connection point. Time series alignment refers to interpolating data from different sources to a unified time point according to timestamps to form a synchronized multi-source time series dataset. Real-time meteorological monitoring data includes the total irradiance and diffuse irradiance of the array tilt surface, the photovoltaic panel backsheet temperature, and the ambient temperature; electrical operation data includes the DC side input voltage, input current, AC side output three-phase voltage, output current, instantaneous output power, and the active and reactive power on the high-voltage side of the transformer substation, as well as the grid connection point voltage and grid connection point current.

[0026] The beneficial effects of the above technical solution are as follows: the actual light-receiving conditions of the photovoltaic module are directly reflected by the tilted surface irradiance measurement; the impact of the module's operating temperature on efficiency is considered by the backsheet temperature acquisition; the power flow is monitored throughout the process by acquiring the AC and DC electrical quantities of the inverter; the grid-side constraints are obtained by the grid-connected status parameters; and the unified time scale ensures the synchronization of multi-source data, providing a high-quality data foundation for subsequent power prediction and scheduling calculations.

[0027] In some embodiments of this application, the power prediction module is used for: Feature extraction is performed on the multi-source time series dataset to obtain trend components and volatility components; The trend component and the fluctuation component are sliced ​​according to the time window. The trend component value corresponding to each prediction time is input into the prediction branch based on the physical model. Combined with the photovoltaic module temperature correction coefficient and line loss, the first power prediction value is generated. The volatility component values ​​corresponding to each prediction time are input into the data-driven prediction branch, and a second power prediction value is generated using historical similar day data. The first power prediction value and the second power prediction value are weighted and fused to obtain the ultra-short-term power prediction sequence within a preset time period and the short-term power prediction sequence within the next few hours.

[0028] In this embodiment, feature extraction employs Empirical Mode Decomposition (EMD) or Wavelet Transform to decompose the multi-source time-series dataset into trend components (slowly changing baselines, such as trends caused by cloud movement) and fluctuation components (high-frequency random fluctuations, such as local shading). The prediction branch based on the physical model uses an equivalent circuit model of the photovoltaic module, inputs the trend component values ​​to obtain the module's ideal output power, and subtracts line losses from the product of the module's ideal output power and the temperature correction coefficient to obtain the first power prediction value. Line losses refer to the energy loss caused by factors such as conductor resistance and transformer impedance during power transmission, and are usually expressed in the form of power. The data-driven prediction branch uses a similar day matching algorithm: it retrieves the date most similar to the current day's meteorological characteristics (season, weather type, irradiance distribution) of the fluctuation component from the historical database, and extracts the actual power curve of that day as the second power prediction value. When generating an ultra-short-term power prediction sequence based on the first and second power prediction values ​​corresponding to the acquisition time, the weight of the first power prediction value is 0.7 and the weight of the second power prediction value is 0.3; when generating a short-term power prediction sequence based on the first and second power prediction values ​​corresponding to the acquisition time, the weight of the first power prediction value is 0.3 and the weight of the second power prediction value is 0.7.

[0029] The beneficial effects of the above technical solution are as follows: it enables the separate processing of different change characteristics through trend-fluctuation decomposition; it utilizes the physical laws of photovoltaic conversion through physical models to ensure the interpretability of predictions; it utilizes the statistical laws of historical experience through data-driven models to adapt to complex meteorological conditions; it combines the advantages of the two types of models through weight fusion; and it adapts to the changes of dominant factors in different forecast periods through time-scale differentiated weights, thereby improving the prediction accuracy across all time scales.

[0030] In some embodiments of this application, the deviation calculation module is used for: Analyze multiple active power dispatch limits issued by the power grid dispatching agency and extract the target output value of the planned processing curve for the corresponding period in the next control cycle; Extract the predicted output value with the same control period from the ultra-short-term power prediction sequence; Calculate the difference between the predicted output value and the target output value, and use it as the basic deviation. Collect real-time evaluation deviations of grid connection points and power deviations of tie lines, and calculate power compensation based on grid frequency regulation requirements; The active power deviation is obtained by vector addition of the basic deviation and the power compensation.

[0031] In this embodiment, a planned processing curve is constructed based on multiple active power dispatch limits. The next control cycle is 15 seconds to 1 minute, and the target output value for the corresponding time period is obtained through interpolation. The time resolution of the ultra-short-term power prediction sequence is 15 seconds, and the predicted output value at the corresponding moment is extracted. The basic deviation is calculated by subtracting the target output value from the predicted output value. A positive value indicates that the predicted over-generation needs to be adjusted downward, and a negative value indicates that the predicted under-generation needs to be adjusted upward. The real-time frequency deviation refers to the difference between the measured frequency at the grid connection point and 50 Hz; the tie-line power deviation refers to the difference between the actual power flow and the planned power flow on the tie-line. The power compensation amount is calculated based on the grid frequency regulation requirements: the primary frequency regulation compensation amount is equal to the frequency deviation multiplied by the frequency regulation coefficient (e.g., 5% per Hz of the rated capacity), and the secondary frequency regulation compensation amount is given by the automatic generation control (AGC) command. Vector superposition refers to the algebraic addition of the basic deviation amount and the power compensation amount, considering the direction (upward or downward adjustment), to generate the total active power deviation amount.

[0032] The beneficial effects of the above technical solution are as follows: the tracking target is clarified by analyzing the dispatch limit; the expected actual output in the future is obtained through ultra-short-term forecasting; the basic deviation quantifies the difference between the plan and the forecast; by introducing frequency deviation and tie-line power deviation, the frequency regulation and power balance needs of the power grid are responded to; the active power deviation generated by vector superposition integrates the dual objectives of plan tracking and power grid support, enabling photovoltaic dispatch to upgrade from simply executing the plan to actively supporting the power grid.

[0033] In some embodiments of this application, the power setting module is used for: Based on the electrical topology of the photovoltaic power station, the inverters are divided into multiple clusters, with each cluster corresponding to a collector line; Calculate the current available power generation margin of each cluster and the line loss coefficient to the grid connection point, and construct a cluster adjustment priority sequence; According to the cluster adjustment priority sequence, the active power deviation is allocated to each cluster in sequence until the deviation is fully allocated or the cluster margin is exhausted. For each cluster that has obtained the deviation allocation, the active power setpoint of a single inverter is calculated based on the current operating status and efficiency characteristic curve of each inverter in the cluster.

[0034] In this embodiment, the electrical topology of a photovoltaic power station is typically as follows: photovoltaic modules → combiner box → DC cable → inverter → AC cable → transformer substation → collector line → step-up substation → grid connection point. Cluster division is based on collector lines. For example, a 50 MW power station with 4 collector lines, each carrying a 12.5 MW inverter, is divided into 4 clusters. The line loss coefficient is calculated by multiplying the line resistance from the cluster to the grid connection point by the square of the current, divided by the voltage. This reflects the line transmission loss; clusters with lower loss coefficients have higher priority (higher regulation efficiency). The cluster regulation priority sequence is ordered from smallest to largest by the loss coefficient, with lower-loss clusters allocated first. Deviation allocation uses an iterative method: the highest-priority cluster first assumes its entire margin, and the remaining deviation is assumed by the next cluster, until the deviation allocation is completed or all cluster margins are exhausted.

[0035] The beneficial effects of the above technical solution are as follows: by dividing large-scale power plants into controllable units through cluster partitioning, the scheduling complexity is reduced; by introducing transmission efficiency factors through line loss coefficients, the cluster adjustment sequence is optimized; by iterative allocation of priority sequences, the efficient absorption of deviations is achieved; and by adopting a hierarchical architecture (power plant-cluster-inverter), both global optimization and local fine control are taken into account, solving the problem of the large number of inverters in large-scale photovoltaic power plants and the difficulty of direct unified scheduling.

[0036] In some embodiments of this application, the power setting module is used for: Obtain the rated capacity and current actual output power of each inverter in the cluster, and calculate the difference to obtain the remaining adjustable capacity of a single unit. The total adjustable margin of the cluster is obtained by counting the number of inverters in grid-connected operation within the cluster and summing the remaining adjustable capacity of each individual unit. The system collects the real-time current carrying capacity and current transmission power of the corresponding power collection lines of the data acquisition cluster, and calculates the remaining transmission capacity of the lines. Compare the total adjustable margin of the cluster with the remaining transmission capacity of the line, and take the minimum of the two as the current available power generation margin.

[0037] In this embodiment, the inverter's rated capacity is the nameplate value (e.g., 500 kW), and the current actual output power is obtained through inverter communication (e.g., 350 kW). The remaining adjustable capacity of a single unit is the rated capacity minus the actual capacity (150 kW). Grid-connected operation status refers to the inverter being in "operation" rather than "standby" or "fault" status, determined by the inverter's status word. The total adjustable margin of the cluster is the sum of the remaining capacities of each grid-connected inverter. For example, if 10 inverters each have 150 kW remaining capacity, the total margin is 1.5 MW. The real-time current carrying capacity of the collector line is obtained through line temperature monitoring or a current carrying capacity calculation model (e.g., current carrying capacity of 400 amps at an ambient temperature of 30°C). The current transmission power is measured on the high-voltage side of the transformer substation (e.g., 10 kV voltage, 200 amp current, power 3.46 MW). The remaining transmission capacity of the line is the rated capacity of the line minus the current power. For example, if the rated capacity is 5 MW, the remaining capacity is 1.54 MW. The smaller value is used for comparison: If the total adjustable margin of the cluster (1.5 MW) is less than the remaining 1.54 MW of the line, then the current available generation margin is 1.5 MW, limited by the inverter capacity; otherwise, it is limited by the line transmission capacity. This value represents the maximum power that the cluster can still increase under the current operating conditions (the downward adjustment margin is calculated similarly).

[0038] The beneficial effects of the above technical solution are: the adjustment space of a single unit is quantified by the difference between the rated and actual power; grid connection status statistics ensure that only available equipment participates in dispatch; transmission constraints are introduced by monitoring line current carrying capacity; and the combined constraints of power generation capacity and transmission capacity are achieved by taking the smaller value, avoiding mismatches such as inverters having margin but lines being overloaded, or lines having margin but inverters being fully loaded, thus ensuring the executability of dispatch instructions.

[0039] In some embodiments of this application, the power setting module is used for: Obtain the maximum power point tracking efficiency and conversion efficiency of each inverter in the cluster under the current irradiance and temperature conditions; The efficiency values ​​are normalized to eliminate dimensional differences and then used as weighting coefficients. The power deviation that the cluster needs to bear is allocated proportionally according to the proportion of each inverter's weight coefficient to the total weight of the cluster. The allocated power regulation is algebraically calculated with the corresponding inverter's current actual output power to generate the active power setpoint, which is then constrained within the inverter's allowable output range.

[0040] In this embodiment, Maximum Power Point Tracking (MPPT) efficiency refers to the inverter's ability to maintain the photovoltaic module's output power at its maximum power point, typically ranging from 95% to 99%; conversion efficiency refers to the efficiency of converting DC to AC, typically ranging from 96% to 98%. The product of the two is the overall efficiency. Efficiency under current irradiance and temperature conditions is obtained through real-time reporting by the inverter or by looking up a table. Normalization involves dividing each inverter's efficiency by the highest efficiency within the cluster to obtain a relative efficiency weight between 0 and 1. Power deviation allocation refers to the allocation of the cluster's required adjustment (e.g., an increase of 1 MW) according to its weight: if an inverter has a weight of 0.12 (highest efficiency), then 0.12 MW (120 kW) is allocated. Algebraic calculation involves adding the allocated adjustment (120 kW increase) to the current actual power (350 kW) to obtain the setpoint of 470 kW. The constraint on the allowable output range means that the set value does not exceed the rated capacity (500 kW) and is not lower than the minimum operating power (e.g., 10% of the rated capacity, i.e., 50 kW). If the calculated value exceeds the limit, the boundary value is taken, and the remaining deviation is borne by other inverters in the cluster.

[0041] The beneficial effects of the above technical solution are as follows: by introducing MPPT efficiency and conversion efficiency, the power generation efficiency of each inverter under the current operating conditions is quantified; by normalizing the weights, efficiency-priority regulation allocation is realized, with high-efficiency inverters undertaking more regulation tasks, thereby improving the overall regulation efficiency; by algebraic operations and boundary constraints, the physical realizability of the set values ​​is ensured; and by the over-limit redistribution mechanism, the deviation is completely absorbed, avoiding the regulation dead zone.

[0042] In some embodiments of this application, the power output module is used for: The active power setpoint is converted into a reference value for the AC side current amplitude of the inverter, and a current vector command is generated by combining the current grid voltage phase information. The current vector command is decomposed into active current components and reactive current components, and encapsulated into a standard remote control message through a communication protocol. The remote control message is transmitted to the target inverter via a fiber optic ring network or power line carrier communication channel; After receiving the message, the inverter analyzes and extracts the active current component, and generates a pulse width modulation signal through proportional-integral operation.

[0043] In this embodiment, the active power setpoint (e.g., 470 kW) is converted into a current amplitude reference value: based on the current grid voltage (e.g., 400 V line voltage), the current amplitude is calculated as power / ( Multiplying by the voltage and power factor, for example, if the power factor is 0.95, the current amplitude is approximately 715 amps. Grid voltage phase information is obtained through a phase-locked loop (PLL) to generate a current vector command (d-axis active, q-axis reactive) in a rotating coordinate system synchronized with the grid (dq coordinate system). Decomposition refers to projecting the vector command onto the d-axis and q-axis to obtain the active current component (id_ref) and reactive current component (iq_ref). Standard remote control messages use ModbusTCP, IEC61850GOOSE, or proprietary protocols, containing fields such as target address, command type, current component value, and checksum. Fiber optic ring network refers to a fiber optic Ethernet network within a power plant (such as a ring network topology composed of industrial Ethernet switches), while power line carrier communication refers to carrier communication (HPLC) that uses power lines to transmit high-frequency signals. After parsing the message, the inverter compares id_ref with the measured id, and through the proportional-integral (PI) controller, generates a duty cycle command for the pulse width modulation (PWM) signal to control the switching timing of the insulated gate bipolar transistor (IGBT), adjust the amplitude and phase of the output current, and make the actual output power track the set value.

[0044] The beneficial effects of the above technical solution are as follows: a bridge between power setting and electrical control is established through the conversion of power to current; active and reactive power decoupling control is achieved through current vector commands; interoperability is ensured through standard communication protocol encapsulation; communication reliability is improved through dual channels of optical fiber and carrier wave; and high-precision power point tracking is achieved through PI control and PWM, completing closed-loop control from scheduling commands to power output.

[0045] In some embodiments of this application, the power correction module is used for: The actual output power value of the inverter after adjustment is collected and compared with the active power set value to calculate the tracking error; The tracking error is correlated with the predicted value of the corresponding time period in the ultra-short-term power prediction sequence to identify the direction of systematic deviation of the prediction model. The prediction deviation coefficient is updated using an exponential smoothing algorithm, so that the corrected prediction value gradually converges to the actual output direction. The updated prediction deviation coefficients are stored in the parameter library for use in the power prediction process of the next control cycle, forming the closed-loop feedback information.

[0046] In this embodiment, the actual output power value is collected by the inverter or transformer substation measurement device (e.g., set at 470 kW, actual output 465 kW), and the tracking error is the actual value minus the set value (-5 kW, negative deviation). Correlation analysis refers to comparing the tracking error with the ultra-short-term predicted value for that period (e.g., predicted 480 kW). If the actual value (465) is less than the predicted value (480), and the tracking error is negative, it is identified as the prediction model being systematically overestimated (the prediction is too optimistic). The exponential smoothing algorithm refers to the deviation coefficient update formula: the new coefficient is equal to the smoothing factor (e.g., 0.3) multiplied by the current error plus (1 minus the smoothing factor) multiplied by the old coefficient, achieving gradual correction rather than abrupt change. The corrected predicted value is equal to the original predicted value multiplied by (1 minus the deviation coefficient), causing the prediction to converge towards the actual output direction. The parameter library is stored in the power plant monitoring system database. The power prediction process in the next control cycle automatically reads the latest coefficients, forming a closed-loop optimization mechanism of "prediction-execution-feedback-correction".

[0047] The beneficial effects of the above technical solution are as follows: quantitative evaluation of control effect is achieved through tracking error acquisition; systematic deviations (too high or too low) of the prediction model are identified through correlation analysis with the predicted value; stable updates of the deviation coefficient are achieved through the exponential smoothing algorithm, avoiding drastic fluctuations; adaptive evolution of the prediction model is achieved through parameter library storage and periodic calls, continuously optimizing the prediction accuracy in long-term operation and solving the problem of photovoltaic power prediction drifting with component aging and environmental changes.

[0048] To further illustrate the technical concept of this invention, the technical solution of this invention will now be described in conjunction with specific application scenarios.

[0049] Correspondingly, such as Figure 2 As shown, this application also provides an active power output scheduling method for a photovoltaic power plant, applied to the active power output scheduling system of a photovoltaic power plant as described in any one of claims 1-9, characterized in that it includes: Real-time meteorological monitoring data and electrical operation data of each photovoltaic array in the photovoltaic power station are obtained to obtain a multi-source time series dataset; Based on the multi-source time-series dataset, a multi-time-scale power prediction process is executed to generate ultra-short-term power prediction sequences and short-term power prediction sequences. Based on the active power dispatch limit issued by the power grid dispatching agency and the ultra-short-term power prediction sequence, calculate the active power deviation for the current control cycle. In response to the active power deviation exceeding the dynamic adjustment dead zone, a hierarchical collaborative scheduling process is triggered to determine the active power setpoint for each inverter unit. The active power setpoint is converted into a modulation command and sent to the corresponding inverter. The actual output power after execution is monitored, and closed-loop feedback information is generated. The prediction deviation coefficient for the next control cycle is corrected based on the closed-loop feedback information, thereby optimizing the subsequent active power output scheduling accuracy.

[0050] In the description of the above embodiments, specific features, structures, materials, or characteristics may be combined in any suitable manner in one or more embodiments or examples.

[0051] Although the invention has been described above with reference to embodiments, various modifications can be made and components can be replaced with equivalents without departing from the scope of the invention. In particular, as long as there is no structural conflict, the features in the embodiments disclosed in this invention can be combined with each other in any way. The fact that not all of these combinations are described in this specification is merely for the sake of brevity and resource conservation.

[0052] It will be understood by those skilled in the art that the above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A photovoltaic power plant active power output dispatching system, characterized in that, include: The collection builds a module to acquire real-time meteorological monitoring data and electrical operation data of each photovoltaic array in the photovoltaic power station, and obtain a multi-source time series dataset; The power prediction module is used to perform a multi-timescale power prediction process based on the multi-source time series dataset to generate ultra-short-term power prediction sequences and short-term power prediction sequences. The deviation calculation module is used to calculate the active power deviation of the current control cycle based on the active power dispatch limit issued by the power grid dispatching agency and the ultra-short-term power prediction sequence. The power setting module is used to trigger the hierarchical collaborative scheduling process and determine the active power setting value of each inverter unit in response to the active power deviation exceeding the dynamic adjustment dead zone. The power output module is used to convert the active power setpoint into a modulation command and send it to the corresponding inverter, monitor the actual output power after execution, and generate closed-loop feedback information. The power correction module is used to correct the prediction deviation coefficient of the next control cycle based on the closed-loop feedback information, thereby optimizing the scheduling accuracy of subsequent active power output.

2. The active power output dispatching system for a photovoltaic power station according to claim 1, characterized in that, The collection building module is used for: The total irradiance and scattered irradiance of the array tilt surface are collected by an irradiance sensor, and the back panel temperature of the photovoltaic panel and the ambient temperature are collected based on a temperature sensor. Collect the DC side input voltage and input current, AC side three-phase output voltage and output current of each inverter, and calculate the instantaneous output power; Record the active power, reactive power, grid connection point voltage, and grid connection point current on the high-voltage side of the transformer substation as grid connection status parameters; The real-time meteorological monitoring data and electrical operation data are time-stamped according to a unified sampling period, and the time series are aligned to obtain the multi-source time series dataset.

3. The active power output dispatching system for a photovoltaic power station according to claim 1, characterized in that, The power prediction module is used for: Feature extraction is performed on the multi-source time-series dataset to obtain trend components and volatility components; The trend component and the fluctuation component are sliced ​​according to the time window. The trend component value corresponding to each prediction time is input into the prediction branch based on the physical model. Combined with the photovoltaic module temperature correction coefficient and line loss, the first power prediction value is generated. The volatility component values ​​corresponding to each prediction time are input into the data-driven prediction branch, and a second power prediction value is generated using historical similar day data. The first power prediction value and the second power prediction value are weighted and fused to obtain the ultra-short-term power prediction sequence within a preset time period and the short-term power prediction sequence within the next few hours.

4. The active power output dispatching system for a photovoltaic power station according to claim 1, characterized in that, The deviation calculation module is used for: Analyze multiple active power dispatch limits issued by the power grid dispatching agency and extract the target output value of the planned processing curve for the corresponding period in the next control cycle; Extract the predicted output value with the same control period from the ultra-short-term power prediction sequence; Calculate the difference between the predicted output value and the target output value, and use it as the basic deviation. Collect real-time evaluation deviations of grid connection points and power deviations of tie lines, and calculate power compensation based on grid frequency regulation requirements; The active power deviation is obtained by vector addition of the basic deviation and the power compensation.

5. The active power output dispatching system for a photovoltaic power station according to claim 1, characterized in that, The power setting module is used for: Based on the electrical topology of the photovoltaic power station, the inverters are divided into multiple clusters, with each cluster corresponding to a collector line; Calculate the current available power generation margin of each cluster and the line loss coefficient to the grid connection point, and construct a cluster adjustment priority sequence; According to the cluster adjustment priority sequence, the active power deviation is allocated to each cluster in sequence until the deviation is fully allocated or the cluster margin is exhausted. For each cluster that has obtained the deviation allocation, the active power setpoint of a single inverter is calculated based on the current operating status and efficiency characteristic curve of each inverter in the cluster.

6. The active power output dispatching system for a photovoltaic power station according to claim 5, characterized in that, The power setting module is used for: Obtain the rated capacity and current actual output power of each inverter in the cluster, and calculate the difference to obtain the remaining adjustable capacity of a single unit. The total adjustable margin of the cluster is obtained by counting the number of inverters in grid-connected operation within the cluster and summing the remaining adjustable capacity of each individual unit. The system collects the real-time current carrying capacity and current transmission power of the corresponding power collection lines of the data acquisition cluster, and calculates the remaining transmission capacity of the lines. Compare the total adjustable margin of the cluster with the remaining transmission capacity of the line, and take the minimum of the two as the current available power generation margin.

7. The active power output dispatching system for a photovoltaic power station according to claim 5, characterized in that, The power setting module is used for: Obtain the maximum power point tracking efficiency and conversion efficiency of each inverter in the cluster under the current irradiance and temperature conditions; The efficiency values ​​are normalized to eliminate dimensional differences and then used as weighting coefficients. The power deviation that the cluster needs to bear is allocated proportionally according to the proportion of each inverter's weight coefficient to the total weight of the cluster. The allocated power regulation is algebraically calculated with the corresponding inverter's current actual output power to generate the active power setpoint, which is then constrained within the inverter's allowable output range.

8. The active power output dispatching system for a photovoltaic power station according to claim 1, characterized in that, The power output module is used for: The active power setpoint is converted into a reference value for the AC side current amplitude of the inverter, and a current vector command is generated by combining the current grid voltage phase information. The current vector command is decomposed into active current components and reactive current components, and encapsulated into a standard remote control message through a communication protocol. The remote control message is transmitted to the target inverter via a fiber optic ring network or power line carrier communication channel; After receiving the message, the inverter analyzes and extracts the active current component, and generates a pulse width modulation signal through proportional-integral operation.

9. The active power output dispatching system for a photovoltaic power station according to claim 1, characterized in that, The power correction module is used for: The actual output power value of the inverter after adjustment is collected and compared with the active power set value to calculate the tracking error; The tracking error is correlated with the predicted value of the corresponding time period in the ultra-short-term power prediction sequence to identify the direction of systematic deviation of the prediction model. The prediction deviation coefficient is updated using an exponential smoothing algorithm, so that the corrected prediction value gradually converges to the actual output direction. The updated prediction deviation coefficients are stored in the parameter library for use in the power prediction process of the next control cycle, forming the closed-loop feedback information.

10. A method for active power output scheduling of a photovoltaic power plant, applied to the active power output scheduling system of the photovoltaic power plant as described in any one of claims 1-9, characterized in that, include: Real-time meteorological monitoring data and electrical operation data of each photovoltaic array in the photovoltaic power station are obtained to obtain a multi-source time series dataset; Based on the multi-source time-series dataset, a multi-time-scale power prediction process is executed to generate ultra-short-term power prediction sequences and short-term power prediction sequences. Based on the active power dispatch limit issued by the power grid dispatching agency and the ultra-short-term power prediction sequence, calculate the active power deviation for the current control cycle. In response to the active power deviation exceeding the dynamic adjustment dead zone, a hierarchical collaborative scheduling process is triggered to determine the active power setpoint for each inverter unit. The active power setpoint is converted into a modulation command and sent to the corresponding inverter. The actual output power after execution is monitored, and closed-loop feedback information is generated. The prediction deviation coefficient for the next control cycle is corrected based on the closed-loop feedback information, thereby optimizing the subsequent active power output scheduling accuracy.

Images