An adaptive grid-connected control method and system for solar photovoltaic power plants

Abstract

This invention relates to the field of solar photovoltaic power plant technology, specifically to an adaptive solar photovoltaic power plant grid-connected control method and system, comprising: acquiring weather forecast data for a first region, historical power generation data of the photovoltaic power plant, and real-time power generation data; training an output power prediction model based on the weather forecast data and historical power generation data, and dynamically updating it using real-time power generation data; predicting future photovoltaic power generation output power in real time using the updated output power prediction model to obtain the predicted photovoltaic power generation output power; analyzing the predicted photovoltaic power generation output power in conjunction with the current grid load demand to obtain current fluctuation trends and load change trends; determining the current fluctuation state through load status and current fluctuation state to obtain the fault current; and constructing a fault current compensation model based on the predicted photovoltaic power generation output power and current fluctuation trends to dynamically compensate for the fault current and adjust the fault current detection strategy of the relay protection device.

Description

Technical Field

[0001] This invention relates to the field of solar photovoltaic power plant technology, specifically to an adaptive solar photovoltaic power plant grid connection control method and system. Background Technology

[0002] With the growth of global energy demand and the development of renewable energy technologies, solar photovoltaic (PV) power generation, as a clean and sustainable energy source, is gradually becoming an important component of the power system. However, the integration of large-scale PV power plants into the distribution network can have certain impacts on the operation of the traditional power grid, especially under low-load or half-load conditions, potentially triggering a series of technical problems. During the grid connection process of PV power generation, due to the intermittent and fluctuating nature of PV power generation, its output power is significantly affected by weather changes. This fluctuation leads to significant changes in the grid current level, especially under low-load conditions, where the current level may drop significantly, making it impossible for relay protection devices to accurately detect fault currents, thereby affecting the safe operation of the power grid.

[0003] However, under low load or abnormal current fluctuation conditions, existing fault current compensation models struggle to respond in real time to photovoltaic power generation fluctuations and current level changes, failing to effectively compensate for errors occurring during fault current detection. Therefore, there is an urgent need for a fault current detection method to accurately and efficiently detect fault currents under low load or abnormal current fluctuation conditions.

[0004] To address this, an adaptive grid-connected control method and system for solar photovoltaic power plants is proposed. Summary of the Invention

[0005] The purpose of this invention is to provide an adaptive grid-connected control method and system for solar photovoltaic power plants, suitable for accurate fault current detection under low load or abnormal current fluctuation conditions. The method includes: acquiring weather forecast data for a first region, historical power generation data of the photovoltaic power plant, and real-time power generation data; training an output power prediction model based on the weather forecast data and historical power generation data, and dynamically updating it using real-time power generation data; predicting future photovoltaic power generation output power in real time using the updated output power prediction model to obtain the predicted photovoltaic power generation output power; analyzing the predicted photovoltaic power generation output power in conjunction with the current grid load demand to obtain current fluctuation trends and load change trends; determining the fault current based on the load status and current fluctuation status; and constructing a fault current compensation model based on the predicted photovoltaic power generation output power and current fluctuation trends to dynamically compensate for the fault current and adjust the fault current detection strategy of the relay protection device.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] An adaptive grid-connected control method for a solar photovoltaic power station includes:

[0008] Acquire weather forecast data and historical power generation data of photovoltaic power plants in the first region; collect real-time power generation data of photovoltaic power plants through intelligent sensors;

[0009] The output power prediction model is trained based on the weather forecast data and the historical power generation data; the output power prediction model is dynamically updated using the real-time power generation data to obtain the updated output power prediction model; the future photovoltaic power generation output power is predicted in real time using the updated output power prediction model to obtain the predicted photovoltaic power generation output power.

[0010] Based on the current load demand of the power grid, the predicted photovoltaic power output is analyzed to obtain the current fluctuation trend and load change trend;

[0011] The current fluctuation state is determined by the load change trend, the predicted photovoltaic power output, and the current fluctuation trend; the specific determination process is as follows:

[0012] The load status is determined based on the load change trend; if the load status is low load and / or half load, and the predicted photovoltaic power output is higher than the expected output, the current fluctuation status is detected as abnormal based on the current fluctuation trend; if the current fluctuation status is abnormal, it is determined to be a fault current.

[0013] Based on the predicted photovoltaic power output and the current fluctuation trend, a fault current compensation model is constructed to dynamically compensate the fault current, and the fault current detection strategy of the relay protection device is adjusted according to expert experience.

[0014] Preferably, the weather forecast data includes: temperature, humidity, solar radiation intensity, wind speed, and cloud cover in the first region;

[0015] The historical power generation data includes: historical output power data, historical current, historical voltage, and historical environmental data of the photovoltaic power station within the specified time period, which are the same as the weather forecast data;

[0016] The real-time power generation data includes: real-time output power, real-time grid load, real-time current, and real-time voltage at the current time.

[0017] Preferably, the output power prediction model includes: a data preprocessing unit, a weather forecast data analysis unit, and a photovoltaic power generation output power initial prediction unit;

[0018] The data preprocessing unit uses statistical analysis to detect and remove outliers in the weather forecast data and historical power generation data; and aligns the weather forecast data and historical power generation data according to time series.

[0019] The weather forecast data analysis unit extracts features related to photovoltaic power generation output from the weather forecast data to obtain weather features; the weather features include: cumulative solar radiation intensity, average daily temperature over the past N days, average daily cloud cover, and average daily humidity;

[0020] The photovoltaic power output power initial prediction unit uses an LSTM model to predict the photovoltaic power output power based on the weather characteristics and the historical power generation data, and obtains the initial predicted photovoltaic power output power.

[0021] Preferably, the output power prediction model is dynamically updated using the real-time power generation data to obtain an updated output power prediction model; the future photovoltaic power generation output power is then predicted in real time using the updated output power prediction model to obtain the predicted photovoltaic power generation output power; the specific process includes:

[0022] The power generation error is calculated by comparing the real-time output power in the real-time power generation data with the initial predicted photovoltaic power generation output power.

[0023] The output power prediction model is updated using a sliding window method based on the power generation error; the update frequency is once every two hours.

[0024] The updated output power prediction model calculates the future photovoltaic power output in real time based on the weather forecast data for the current time and future periods, thus obtaining the predicted photovoltaic power output.

[0025] Preferably, the specific process for determining the load status based on the load change trend is as follows: calculate the load deviation based on the current power grid load demand and the historical load average; perform a first-order derivative on the load change trend to obtain the load change rate; and set a load status threshold based on the load deviation and the load change rate to determine the load status.

[0026] Preferably, the specific process for detecting whether the current fluctuation is in an abnormal fluctuation state based on the current fluctuation trend is as follows: the average amplitude of the real-time current fluctuation is calculated from the real-time current in the real-time power generation data; the average amplitude of the historical current fluctuation is calculated from the historical current; by comparing the average amplitude of the real-time current fluctuation and the average amplitude of the historical current fluctuation, it is determined whether the current fluctuation exceeds a preset normal range; if the current fluctuation exceeds the preset normal range, it is determined to be an abnormal fluctuation state.

[0027] Preferably, the fault current compensation model includes: a fault current data input layer, a dynamic compensation layer, and a relay protection strategy adjustment layer;

[0028] The fault current data input layer inputs the predicted photovoltaic power output, the current fluctuation trend, and the current deviation into the fault current compensation model; wherein, the current deviation is obtained by calculating the difference between the real-time current and the predicted current;

[0029] The dynamic compensation layer calculates the fault current compensation amount based on the predicted photovoltaic power output, the current fluctuation trend, and the current deviation.

[0030] The relay protection strategy adjustment layer adjusts the sensitivity of the relay protection device and the fault current detection threshold according to the fault current compensation amount.

[0031] Preferably, the formula for calculating the fault current compensation is:

[0032] △I=k1·f(P)+k2·△I volatility +k3·△I offset ;

[0033]

[0034] Where △I is the fault current compensation amount; k1 is the equivalent current compensation coefficient for the output power difference; f(P) is the function that transforms the output power difference into an equivalent circuit; P is the output power difference; k2 is the current fluctuation amplitude compensation coefficient; △I volatility k3 is the current fluctuation amplitude; k3 is the current deviation compensation coefficient; ΔI offset For current deviation; P forecast To predict the output power of photovoltaic power generation; P actual V represents the real-time output power of photovoltaic power generation. actual α is the real-time voltage; I is the real-time voltage adjustment coefficient; actual This represents the real-time current.

[0035] An adaptive grid-connected control system for a solar photovoltaic power station includes:

[0036] The data acquisition module is used to acquire weather forecast data and historical power generation data of the photovoltaic power station in the first region; and to collect real-time power generation data of the photovoltaic power station through intelligent sensors.

[0037] The output power prediction module is used to train an output power prediction model based on the weather forecast data and the historical power generation data; dynamically update the output power prediction model using the real-time power generation data to obtain the updated output power prediction model; and make real-time predictions of future photovoltaic power generation output power using the updated output power prediction model to obtain the predicted photovoltaic power generation output power.

[0038] The output power analysis module is used to analyze the predicted photovoltaic power output power in conjunction with the current grid load demand, and obtain the current fluctuation trend and load change trend.

[0039] The fault current determination module is used to determine the current fluctuation state based on the load change trend, the predicted photovoltaic power output, and the current fluctuation trend; the specific determination process is as follows:

[0040] The load status is determined based on the load change trend; if the load status is low load and / or half load, and the predicted photovoltaic power output is higher than the expected output, the current fluctuation status is detected as abnormal based on the current fluctuation trend; if the current fluctuation status is abnormal, it is determined to be a fault current.

[0041] The fault current detection strategy adjustment module is used to construct a fault current compensation model based on the predicted photovoltaic power output and the current fluctuation trend to dynamically compensate the fault current, and adjust the fault current detection strategy of the relay protection device according to expert experience.

[0042] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0043] 1. This invention proposes an output power prediction model for real-time prediction of future photovoltaic power output. By acquiring weather forecast data, historical power generation data, and real-time power generation data, the output power prediction model is dynamically constructed and updated, which can significantly improve the accuracy of photovoltaic power generation prediction and enable the system to quickly adapt to environmental changes. Real-time prediction of future power output assists the power grid in precise load management and optimized scheduling, reducing the impact of grid connection fluctuations on power grid operation. It also provides data support for the adjustment of relay protection sensitivity, improves the accuracy of fault detection under low load or abnormal current fluctuation conditions, and effectively solves the problem of inaccurate power output prediction caused by the strong volatility of photovoltaic power generation, providing more stable and reliable technical support for power grid operation.

[0044] 2. This invention proposes a fault current detection method based on load change trends and photovoltaic power output analysis. This method accurately determines the grid load status by real-time analysis of load change trends and predicted photovoltaic power output, and further determines whether the current is in an abnormal fluctuation state by combining current fluctuation trends. This analysis based on predicted power and load changes helps to accurately identify fault currents, avoid false alarms and missed alarms caused by load changes, improve the accuracy of fault current detection and system response speed, and provide efficient and reliable fault current detection support for subsequent dynamic compensation of fault currents through fault current compensation models.

[0045] 3. This invention proposes a fault current compensation model based on predicted photovoltaic power output and current fluctuation trends, and dynamically adjusts the fault current detection strategy of the relay protection device, which helps to achieve more accurate and intelligent fault current detection. This method can automatically adjust the sensitivity of the relay protection system according to the predicted photovoltaic power output and current fluctuation trends to cope with current changes caused by photovoltaic power fluctuations, ensuring accurate detection of fault currents even under low or high load conditions. This dynamic compensation mechanism can reduce the risk of malfunction and missed operation of the relay protection system in photovoltaic power plants, improve system reliability, avoid unnecessary protection actions or undetectable faults caused by current fluctuations, thereby ensuring the stable grid-connected operation of photovoltaic power plants and optimizing the overall safety and stability of the power grid. Attached Figure Description

[0046] Figure 1 A flowchart of an adaptive solar photovoltaic power station grid connection control method provided in an embodiment of the present invention;

[0047] Figure 2 A structural diagram of an adaptive solar photovoltaic power station grid-connected control system provided in an embodiment of the present invention;

[0048] Figure 3 A schematic diagram of the fault current determination process provided in an embodiment of the present invention;

[0049] Figure 4 This is a schematic diagram of dynamic fault current compensation provided in an embodiment of the present invention. Detailed Implementation

[0050] 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.

[0051] With the growth of global energy demand and the development of renewable energy technologies, solar photovoltaic (PV) power generation, as a clean and sustainable energy source, is gradually becoming an important component of the power system. However, the integration of large-scale PV power plants into the distribution network can have certain impacts on the operation of the traditional power grid, especially under low-load or half-load conditions, potentially triggering a series of technical problems. During the grid connection process of PV power generation, due to the intermittent and fluctuating nature of PV power generation, its output power is significantly affected by weather changes. This fluctuation leads to significant changes in the grid current level, especially under low-load conditions, where the current level may drop significantly, making it impossible for relay protection devices to accurately detect fault currents, thereby affecting the safe operation of the power grid.

[0052] This invention proposes an adaptive grid-connected control method for solar photovoltaic power plants, achieving efficient and accurate detection and dynamic compensation of fault currents under low load conditions. This method is applied to an adaptive grid-connected control system for solar photovoltaic power plants. Detailed flowcharts and system structure diagrams are provided below. Figure 1 and Figure 2 To illustrate that the method and system of the present invention can achieve efficient and accurate detection and dynamic compensation of fault current, the effectiveness of the present invention will be explained below through two embodiments.

[0053] Example 1

[0054] In this application embodiment, the method and system proposed in this invention are used to describe in detail the efficient and accurate detection and dynamic compensation process of fault current. In this application embodiment, the efficient and accurate detection and dynamic compensation of fault current is specifically addressed to the efficient and accurate detection and dynamic compensation of fault current in the relay system of a solar photovoltaic power station (station A) in a certain region. The following is based on... Figure 1 and Figure 2 The content provides a detailed explanation of the efficient and accurate detection and dynamic compensation process for fault current in the relay system of this solar photovoltaic power station; among other things, Figure 1The specific process of the method proposed in this invention includes: acquiring weather forecast data and historical power generation data of a photovoltaic power station in a first region; collecting real-time power generation data of the photovoltaic power station through intelligent sensors; training an output power prediction model based on the weather forecast data and the historical power generation data; dynamically updating the output power prediction model using the real-time power generation data to obtain an updated output power prediction model; predicting the future photovoltaic power generation output power in real time using the updated output power prediction model to obtain the predicted photovoltaic power generation output power; analyzing the predicted photovoltaic power generation output power in conjunction with the current grid load demand to obtain current fluctuation trends and load change trends; determining the fault current by combining the load change trend, the predicted photovoltaic power generation output power, and the current fluctuation trend; constructing a fault current compensation model based on the predicted photovoltaic power generation output power and the current fluctuation trend to dynamically compensate the fault current, and adjusting the fault current detection strategy of the relay protection device based on expert experience. Figure 2 The system structure diagram proposed in this invention includes: a data acquisition module, an output power prediction module, an output power analysis module, a fault current determination module, and a fault current detection strategy adjustment module; combined with Figure 1 and Figure 2 The following explanation is provided regarding the content:

[0055] Acquire weather forecast data and historical power generation data of photovoltaic power plants in the first region; collect real-time power generation data of photovoltaic power plants through intelligent sensors;

[0056] The weather forecast data includes: temperature, humidity, solar radiation intensity, wind speed, and cloud cover in the first region;

[0057] The historical power generation data includes: historical output power data, historical current, historical voltage, and historical environmental data of the photovoltaic power station within a specified time period; in this embodiment, historical power generation data of the photovoltaic power station from 2020 to 2022 are obtained.

[0058] The real-time power generation data includes: real-time output power, real-time grid load, real-time current, and real-time voltage at the current time.

[0059] In this embodiment, the weather forecast data, the historical power generation data, and the real-time power generation data are obtained from the power company and meteorological station in the first region.

[0060] In this embodiment, by acquiring weather forecast data for the first region, historical power generation data of the photovoltaic power station, and real-time power generation data, comprehensive data support is provided for accurate prediction of photovoltaic power output and grid load management. Specifically, weather forecast data helps predict changes in the external environment of photovoltaic power generation, thereby affecting the power generation efficiency of photovoltaic modules; historical power generation data provides a long-term reference for model training and power prediction, improving the accuracy and reliability of predictions; real-time power generation data reflects the current operating status of the photovoltaic power station and changes in grid load, supporting real-time monitoring and adjustment. By comprehensively using these data, a reliable data foundation is provided for the subsequent dynamic updating of the output power prediction model to predict future photovoltaic power output.

[0061] Preferably, the output power prediction model is trained based on the weather forecast data and the historical power generation data;

[0062] The output power prediction model includes: a data preprocessing unit, a weather forecast data analysis unit, and a photovoltaic power generation output power initial prediction unit;

[0063] The data preprocessing unit uses statistical analysis to detect and remove outliers in the weather forecast data and historical power generation data; and aligns the weather forecast data and historical power generation data according to time series.

[0064] The weather forecast data analysis unit extracts features related to photovoltaic power generation output from the weather forecast data to obtain weather features; the weather features include: cumulative solar radiation intensity, average daily temperature over the past N days, average daily cloud cover, and average daily humidity;

[0065] The photovoltaic power output power initial prediction unit uses an LSTM model to predict the photovoltaic power output power based on the weather characteristics and the historical power generation data, and obtains the initial predicted photovoltaic power output power.

[0066] In this embodiment, a high-precision photovoltaic (PV) power output prediction model is achieved by combining weather forecast data and historical power generation data and training an output power prediction model using an LSTM model. Data preprocessing removes outliers and performs time series alignment, extracting weather features related to power generation to further improve prediction accuracy. This method can accurately predict future power generation, providing reliable support for grid load management and dispatch, optimizing PV grid-connected control, and improving system stability and energy utilization efficiency.

[0067] Preferably, the output power prediction model is dynamically updated using the real-time power generation data to obtain an updated output power prediction model; the future photovoltaic power generation output power is then predicted in real time using the updated output power prediction model to obtain the predicted photovoltaic power generation output power; the specific process includes:

[0068] The power generation error is calculated by comparing the real-time output power in the real-time power generation data with the initial predicted photovoltaic power generation output power.

[0069] The output power prediction model is updated using a sliding window method based on the power generation error; the update frequency is once every two hours.

[0070] The updated output power prediction model calculates the future photovoltaic power output in real time based on the weather forecast data for the current time and future periods, thus obtaining the predicted photovoltaic power output.

[0071] To further illustrate the advantages of the above LSTM model in predicting photovoltaic power output, Table 1 below shows the comparison results of the present invention with other prediction methods in terms of prediction accuracy and computational performance.

[0072] Table 1 Comparison of Performance Indicators of Photovoltaic Power Generation Output Prediction Results

[0073] method MAE RMSE Predicted time (s) This invention 3.25 12.79 0.35 Regression analysis 5.32 15.20 1.20 CNN 5.71 18.37 0.83

[0074] In this embodiment, the accuracy and real-time performance of the photovoltaic (PV) power output prediction model are improved by dynamically updating the model. By comparing real-time power generation data with the initial predicted power, the power generation error is calculated, and the prediction model is updated using a sliding window method based on this error, enabling the model to adapt to changes in the environment and system in a timely manner. The power output prediction model is updated every two hours, ensuring the real-time accuracy of PV power output prediction. The updated model, combined with weather forecast data, predicts future PV power output in real time, providing precise decision support for grid dispatch and energy management, and effectively improving the regulation capability and grid connection stability of PV power generation.

[0075] Preferably, the predicted photovoltaic power output is analyzed in conjunction with the current grid load demand to obtain the current fluctuation trend and load change trend;

[0076] The current fluctuation state is determined by combining the load change trend, the predicted photovoltaic power output, and the current fluctuation trend, with reference to... Figure 3 The specific determination process is as follows:

[0077] The load status is determined based on the load change trend. The specific process is as follows: the load deviation is calculated by comparing the current power grid load demand with the historical load average; the first derivative of the load change trend is performed to obtain the load change rate; and a load status threshold is set based on the load deviation and the load change rate to determine the load status.

[0078] In this embodiment, by combining load deviation with load change rate, the power grid load status can be accurately determined, avoiding misjudgments caused by a single real-time load value, while adapting to load characteristics under different operating scenarios. By setting a load status threshold, normal fluctuations and abnormal fluctuations can be effectively distinguished, providing a reliable basis for subsequent current fluctuation analysis and fault detection.

[0079] Preferably, if the load state is a low load and / or half load state, and the predicted photovoltaic power output is higher than the expected output power, the current fluctuation state is detected as an abnormal fluctuation state based on the current fluctuation trend; if the current fluctuation state is the abnormal fluctuation state, it is determined to be the fault current.

[0080] The specific process for detecting whether the current fluctuation is in an abnormal fluctuation state based on the current fluctuation trend is as follows: the average amplitude of the real-time current fluctuation is calculated from the real-time current in the real-time power generation data; the average amplitude of the historical current fluctuation is calculated from the historical current; by comparing the average amplitude of the real-time current fluctuation and the average amplitude of the historical current fluctuation, it is determined whether the current fluctuation exceeds the preset normal range; if the current fluctuation exceeds the preset normal range, it is determined to be an abnormal fluctuation state.

[0081] This application's embodiments accurately determine abnormal current fluctuation states by comparing and analyzing real-time current fluctuation amplitudes with historical current fluctuation amplitudes, ensuring accurate identification of fault currents. By dynamically adjusting the detection sensitivity based on different load conditions and current fluctuation characteristics, false positives and false negatives are effectively avoided, improving the system's response speed and adaptability to abnormal fluctuations. This method enhances the intelligent decision-making capabilities of the photovoltaic power station's grid-connected control system, providing reliable support for relay protection optimization under low-load or half-load operating conditions.

[0082] Preferably, based on the predicted photovoltaic power output and the current fluctuation trend, a fault current compensation model is constructed to dynamically compensate the fault current, and the fault current detection strategy of the relay protection device is adjusted according to expert experience. (See reference...) Figure 4 ;

[0083] The fault current compensation model includes: a fault current data input layer, a dynamic compensation layer, and a relay protection strategy adjustment layer.

[0084] The fault current data input layer inputs the predicted photovoltaic power output, the current fluctuation trend, and the current deviation; wherein, the current deviation is obtained by calculating the difference between the real-time current and the predicted current;

[0085] The dynamic compensation layer calculates the fault current compensation amount based on the predicted photovoltaic power output, the current fluctuation trend, and the current deviation.

[0086] The relay protection strategy adjustment layer adjusts the sensitivity of the relay protection device and the fault current detection threshold according to the fault current compensation amount.

[0087] To further illustrate the advantages of the fault current determination method of the present invention in terms of accuracy and false alarm rate, Table 2 provides a comparison of the performance of the present invention with fault current determination methods with different parameters.

[0088] Table 2 Performance Comparison of Fault Current Determination Methods

[0089]

[0090] This application presents a fault current compensation model that dynamically compensates for potential fault currents in the power grid based on the predicted output power of photovoltaic power generation and current fluctuation trends. The model receives and integrates the predicted output power of photovoltaic power generation, current fluctuation trends, and current deviations through a fault current data input layer; a dynamic compensation layer calculates the fault current compensation amount based on these inputs; and a relay protection strategy adjustment layer adjusts the sensitivity of the relay protection device and the fault current detection threshold based on the compensation amount. This dynamic compensation mechanism can adjust the relay protection strategy in real time, avoiding malfunctions or missed detections in the relay protection system caused by current fluctuations, improving the stability and security of the power grid, ensuring the normal operation of photovoltaic power plants, and protecting the power system from faults.

[0091] Preferably, the formula for calculating the fault current compensation is:

[0092] △I=k1·f(P)+k2·△I volatility +k3·△I offset ;

[0093]

[0094] Where △I is the fault current compensation amount; k1 is the equivalent current compensation coefficient for the output power difference; f(P) is the function that transforms the output power difference into an equivalent circuit; P is the output power difference; k2 is the current fluctuation amplitude compensation coefficient; △I volatility k3 is the current fluctuation amplitude; k3 is the current deviation compensation coefficient; ΔI offset For current deviation; P forecastTo predict the output power of photovoltaic power generation; P actual V represents the real-time output power of photovoltaic power generation. actual α is the real-time voltage; I is the real-time voltage adjustment coefficient; actual This represents the real-time current.

[0095] The calculation formula for fault current compensation provided in this application dynamically adjusts the current based on the predicted output power of photovoltaic power generation, current fluctuation trend, and current deviation, ensuring the stability of the power grid during load fluctuations and changes in photovoltaic power generation, and providing a scientific basis for subsequent adjustments to the fault current detection strategy of relay protection devices; by weighting and dynamically adjusting different key parameters, the dynamic compensation process of fault current has higher adaptability.

[0096] This application's embodiments construct a dynamically updated output power prediction model based on weather forecast data, historical power generation data, and real-time power generation data. Combined with grid load demand, it comprehensively analyzes current fluctuation trends and load change trends, achieving accurate detection of abnormal currents under low-load or half-load conditions. Furthermore, by constructing a fault current compensation model, it dynamically compensates for current level changes caused by photovoltaic power generation fluctuations, adjusting the fault current detection strategy of the relay protection device, thereby significantly improving the accuracy and sensitivity of fault detection. This method can adapt to complex and changing operating environments, especially under low-load or half-load operating conditions, effectively avoiding missed or false detections of fault currents by the relay protection system due to current level fluctuations, thus improving the safety and stability of photovoltaic power station grid-connected operation. In addition, the method of this invention relies on intelligent data-driven analysis, reducing human intervention, optimizing system automation performance, and demonstrating highly intelligent grid-connected control capabilities. This ensures that the impact of photovoltaic power stations on grid operation is minimized when connected to the grid, achieving accurate and efficient detection of fault currents under low-load or abnormal current fluctuation conditions.

[0097] Example 2

[0098] In Example 1, the method proposed in this invention successfully achieved efficient and accurate detection and dynamic compensation of fault current in the relay system of a solar photovoltaic power station in region A. To further verify the effectiveness of this invention, this application also proposes an adaptive solar photovoltaic power station grid-connected control system for detecting and compensating for fault current in the relay system of a solar photovoltaic power station in region B; the system includes: a data acquisition module, an output power prediction module, a fault current determination module, and a fault current detection strategy adjustment module.

[0099] The data acquisition module is used to acquire weather forecast data and historical power generation data of the photovoltaic power station in the first region; and to collect real-time power generation data of the photovoltaic power station through intelligent sensors.

[0100] The weather forecast data includes: temperature, humidity, solar radiation intensity, wind speed, and cloud cover in the first region;

[0101] The historical power generation data includes: historical output power data, historical current, historical voltage, and historical environmental data of the photovoltaic power station within a specified time period; in this embodiment, historical power generation data of the photovoltaic power station from 2020 to 2022 are obtained.

[0102] The real-time power generation data includes: real-time output power, real-time grid load, real-time current, and real-time voltage at the current time.

[0103] Preferably, the output power prediction module is used to train an output power prediction model based on the weather forecast data and the historical power generation data;

[0104] The output power prediction model includes: a data preprocessing unit, a weather forecast data analysis unit, and a photovoltaic power generation output power initial prediction unit;

[0105] The data preprocessing unit uses statistical analysis to detect and remove outliers in the weather forecast data and historical power generation data; and aligns the weather forecast data and historical power generation data according to time series.

[0106] The weather forecast data analysis unit extracts features related to photovoltaic power generation output from the weather forecast data to obtain weather features; the weather features include: cumulative solar radiation intensity, average daily temperature over the past N days, average daily cloud cover, and average daily humidity;

[0107] The photovoltaic power output power initial prediction unit uses an LSTM model to predict the photovoltaic power output power based on the weather characteristics and the historical power generation data, and obtains the initial predicted photovoltaic power output power.

[0108] Preferably, the output power prediction model is dynamically updated using the real-time power generation data to obtain an updated output power prediction model; the future photovoltaic power generation output power is then predicted in real time using the updated output power prediction model to obtain the predicted photovoltaic power generation output power; the specific process includes:

[0109] The power generation error is calculated by comparing the real-time output power in the real-time power generation data with the initial predicted photovoltaic power generation output power.

[0110] The output power prediction model is updated using a sliding window method based on the power generation error; the update frequency is once every two hours.

[0111] The updated output power prediction model calculates the future photovoltaic power output in real time based on the weather forecast data for the current time and future periods, thus obtaining the predicted photovoltaic power output.

[0112] Preferably, the output power analysis module is used to analyze the predicted photovoltaic power output power in conjunction with the current grid load demand, and obtain the current fluctuation trend and load change trend.

[0113] Preferably, the fault current determination module is used to determine the current fluctuation state based on the load change trend, the predicted photovoltaic power output, and the current fluctuation trend; the specific determination process is as follows:

[0114] Determine the load status based on the described load change trend;

[0115] If the load condition is low load and / or half load, and the predicted photovoltaic power output is higher than the expected output, the current fluctuation condition is detected as an abnormal fluctuation condition based on the current fluctuation trend.

[0116] If the current fluctuation state is the abnormal fluctuation state, it is determined to be a fault current.

[0117] The specific process for determining the load status based on the load change trend is as follows:

[0118] The load deviation is calculated based on the current power grid load demand and the historical load average.

[0119] The first derivative of the load change trend yields the load change rate.

[0120] Based on the load deviation and the load change rate, a load status threshold is set to determine the load status;

[0121] The specific process for detecting whether the current fluctuation state is in an abnormal fluctuation state based on the current fluctuation trend is as follows:

[0122] The average value of real-time current fluctuation is calculated from the real-time current in the real-time power generation data.

[0123] The average historical current fluctuation range is calculated based on historical current data.

[0124] By comparing the average real-time current fluctuation amplitude with the average historical current fluctuation amplitude, it is determined whether the current fluctuation exceeds the preset normal range.

[0125] If the current fluctuation exceeds the preset normal range, it is determined to be an abnormal fluctuation state.

[0126] Preferably, the fault current detection strategy adjustment module is used to construct a fault current compensation model based on the predicted photovoltaic power output and the current fluctuation trend to dynamically compensate the fault current, and adjust the fault current detection strategy of the relay protection device according to expert experience.

[0127] The fault current compensation model includes: a fault current data input layer, a dynamic compensation layer, and a relay protection strategy adjustment layer.

[0128] The fault current data input layer inputs the predicted photovoltaic power output, the current fluctuation trend, and the current deviation; wherein, the current deviation is obtained by calculating the difference between the real-time current and the predicted current;

[0129] The dynamic compensation layer calculates the fault current compensation amount based on the predicted photovoltaic power output, the current fluctuation trend, and the current deviation; the calculation formula for the fault current compensation amount is:

[0130] △I=k1·f(P)+k2·△I volatility +k3·△I offset ;

[0131]

[0132] Where △I is the fault current compensation amount; k1 is the equivalent current compensation coefficient for the output power difference; f(P) is the function that transforms the output power difference into an equivalent circuit; P is the output power difference; k2 is the current fluctuation amplitude compensation coefficient; △I volatility k3 is the current fluctuation amplitude; k3 is the current deviation compensation coefficient; ΔI offset For current deviation; P forecast To predict the output power of photovoltaic power generation; P actual V represents the real-time output power of photovoltaic power generation. actual α is the real-time voltage; I is the real-time voltage adjustment coefficient; actual For real-time current;

[0133] The relay protection strategy adjustment layer adjusts the sensitivity of the relay protection device and the fault current detection threshold based on expert experience according to the fault current compensation amount.

[0134] Table 3 shows the results of adjusting the sensitivity of the relay protection device and the fault current detection threshold according to the fault current compensation amount.

[0135] Table 3. Results of Adjustment of Sensitivity and Fault Current Detection Threshold of Relay Protection Devices

[0136]

[0137] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An adaptive grid-connected control method for a solar photovoltaic power station, characterized in that, include: Acquire weather forecast data and historical power generation data of photovoltaic power plants in the first region; collect real-time power generation data of photovoltaic power plants through intelligent sensors; The output power prediction model is trained based on the weather forecast data and the historical power generation data; the output power prediction model is dynamically updated using the real-time power generation data to obtain the updated output power prediction model; the future photovoltaic power generation output power is predicted in real time using the updated output power prediction model to obtain the predicted photovoltaic power generation output power. Based on the current load demand of the power grid, the predicted photovoltaic power output is analyzed to obtain the current fluctuation trend and load change trend; The current fluctuation state is determined by the load change trend, the predicted photovoltaic power output, and the current fluctuation trend; the specific determination process is as follows: The load status is determined based on the load change trend; if the load status is low load and / or half load, and the predicted photovoltaic power output is higher than the expected output, the current fluctuation status is detected as an abnormal fluctuation status based on the current fluctuation trend. If the current fluctuation state is the abnormal fluctuation state, it is determined to be a fault current; Based on the predicted photovoltaic power output and the current fluctuation trend, a fault current compensation model is constructed to dynamically compensate the fault current, and the fault current detection strategy of the relay protection device is adjusted according to expert experience.

2. The adaptive solar photovoltaic power station grid-connected control method according to claim 1, characterized in that: The weather forecast data includes: temperature, humidity, solar radiation intensity, wind speed, and cloud cover in the first region; The historical power generation data includes: historical output power data, historical current, historical voltage, and historical environmental data of the photovoltaic power station within a specified time period; The real-time power generation data includes: real-time output power, real-time grid load, real-time current, and real-time voltage at the current time.

3. The adaptive solar photovoltaic power station grid-connected control method according to claim 1, characterized in that: The output power prediction model includes: a data preprocessing unit, a weather forecast data analysis unit, and a photovoltaic power generation output power initial prediction unit; The data preprocessing unit uses statistical analysis to detect and remove outliers in the weather forecast data and historical power generation data; and aligns the weather forecast data and historical power generation data according to time series. The weather forecast data analysis unit extracts features related to photovoltaic power generation output from the weather forecast data to obtain weather features; the weather features include: cumulative solar radiation intensity, average daily temperature over the past N days, average daily cloud cover, and average daily humidity; The photovoltaic power output power initial prediction unit uses an LSTM model to predict the photovoltaic power output power based on the weather characteristics and the historical power generation data, and obtains the initial predicted photovoltaic power output power.

4. The adaptive solar photovoltaic power station grid-connected control method according to claim 1, characterized in that: The output power prediction model is dynamically updated using the real-time power generation data to obtain an updated output power prediction model; the future photovoltaic power generation output power is then predicted in real time using the updated output power prediction model to obtain the predicted photovoltaic power generation output power; the specific process includes: The power generation error is calculated by comparing the real-time output power in the real-time power generation data with the initial predicted photovoltaic power generation output power. The output power prediction model is updated using a sliding window method based on the power generation error; the update frequency is once every two hours. The updated output power prediction model calculates the future photovoltaic power output in real time based on the weather forecast data for the current time and future periods, thus obtaining the predicted photovoltaic power output.

5. The adaptive solar photovoltaic power station grid-connected control method according to claim 1, characterized in that: The specific process for determining the load status based on the load change trend is as follows: The load deviation is calculated by comparing the current load demand of the power grid with the historical average load. The first derivative of the load change trend yields the load change rate. Based on the load deviation and the load change rate, a load status threshold is set to determine the load status.

6. The adaptive solar photovoltaic power station grid-connected control method according to claim 1, characterized in that: The specific process for detecting whether the current fluctuation state is in an abnormal fluctuation state based on the current fluctuation trend is as follows: The average value of real-time current fluctuation is calculated from the real-time current in the real-time power generation data. The average historical current fluctuation range is calculated based on historical current data. By comparing the average real-time current fluctuation amplitude with the average historical current fluctuation amplitude, it is determined whether the current fluctuation exceeds the preset normal range. If the current fluctuation exceeds the preset normal range, it is determined to be an abnormal fluctuation state.

7. The adaptive solar photovoltaic power station grid-connected control method according to claim 1, characterized in that: The fault current compensation model includes: a fault current data input layer, a dynamic compensation layer, and a relay protection strategy adjustment layer. The fault current data input layer inputs the predicted photovoltaic power output, the current fluctuation trend, and the current deviation; wherein, the current deviation is obtained by calculating the difference between the real-time current and the predicted current; The dynamic compensation layer calculates the fault current compensation amount based on the predicted photovoltaic power output, the current fluctuation trend, and the current deviation. The relay protection strategy adjustment layer adjusts the sensitivity of the relay protection device and the fault current detection threshold according to the fault current compensation amount.

8. The adaptive solar photovoltaic power station grid-connected control method according to claim 7, characterized in that: The formula for calculating the fault current compensation is: △I=k1·f(P)+k2·△I volatility +k3·△I offset ; Where △I is the fault current compensation amount; k1 is the equivalent current compensation coefficient for the output power difference; f(P) is the function that transforms the output power difference into an equivalent circuit; P is the output power difference; k2 is the current fluctuation amplitude compensation coefficient; △I volatility k3 is the current fluctuation amplitude; k3 is the current deviation compensation coefficient; ΔI offset For current deviation; P forecast To predict the output power of photovoltaic power generation; P actual V represents the real-time output power of photovoltaic power generation. actual α is the real-time voltage; I is the real-time voltage adjustment coefficient; actual This represents the real-time current.

9. An adaptive grid-connected control system for a solar photovoltaic power station, characterized in that, include: The data acquisition module is used to acquire weather forecast data and historical power generation data of the photovoltaic power station in the first region; and to collect real-time power generation data of the photovoltaic power station through intelligent sensors. The output power prediction module is used to train an output power prediction model based on the weather forecast data and the historical power generation data; dynamically update the output power prediction model using the real-time power generation data to obtain the updated output power prediction model; and make real-time predictions of future photovoltaic power generation output power using the updated output power prediction model to obtain the predicted photovoltaic power generation output power. The output power analysis module is used to analyze the predicted photovoltaic power output power in conjunction with the current grid load demand, and obtain the current fluctuation trend and load change trend. The fault current determination module is used to determine the current fluctuation state based on the load change trend, the predicted photovoltaic power output, and the current fluctuation trend; the specific determination process is as follows: The load status is determined based on the load change trend; if the load status is low load and / or half load, and the predicted photovoltaic power output is higher than the expected output, the current fluctuation status is detected as abnormal based on the current fluctuation trend; if the current fluctuation status is abnormal, it is determined to be a fault current. The fault current detection strategy adjustment module is used to construct a fault current compensation model based on the predicted photovoltaic power output and the current fluctuation trend to dynamically compensate the fault current, and adjust the fault current detection strategy of the relay protection device according to expert experience.

Images