Power distribution network protection device setting method and device based on new energy output prediction, electronic equipment and storage medium

Abstract

The invention discloses a power distribution network protection device setting method and device based on new energy output prediction, electronic equipment and a storage medium, and belongs to the field of intelligent power grids, and the method comprises the steps: collecting the real-time power generation power, the start-stop state, the power plant side environment data and the associated meteorological station environment data of a new energy unit; the power plant side environment data comprises wind speed and wind direction of a wind turbine unit and illumination intensity of a photovoltaic unit. Inputting the wind power related data into the wind power output prediction model to generate future wind power output; and inputting the photovoltaic related data into the photovoltaic output prediction model to generate future photovoltaic output. And according to the wind power and photovoltaic output prediction result, calculating the total new energy power generation amount in the future time period. And determining a setting value of the power distribution network protection device according to the proportion of the predicted generating capacity to the theoretical maximum output of the new energy unit, and performing setting according to the setting value. Therefore, by implementing the invention, the problem of insufficient setting accuracy of the power distribution network protection device in the prior art can be solved.

Description

Technical Field

[0001] The present invention relates to the technical field of smart grids, and particularly relates to a setting method, device, electronic device and storage medium for a distribution network protection device based on new energy output prediction. Background Art

[0002] In a distribution network, the setting value of a protection device is directly related to the accuracy of fault identification and the stability of grid operation. Reasonable setting can quickly isolate the fault area when a fault occurs, prevent the expansion of the accident, and ensure the continuity and safety of power supply. Especially in the context of the increasing proportion of new energy grid connection, due to the intermittency and uncertainty of wind power and photovoltaic power generation, the operation state of the power system is more complex, posing higher requirements for the setting of protection devices. If the setting value is set improperly, it is easy to cause misoperation or refusal to operate of the protection device, resulting in power supply interruption, equipment damage and even grid disturbance.

[0003] However, there are still many deficiencies in the existing technology during the setting calculation process. On the one hand, most setting strategies mainly rely on a single type of historical operation data, lacking the ability to fuse and model multi-dimensional factors such as environmental data, equipment status information and meteorological changes, resulting in deviations in prediction results, thus affecting the accuracy of setting values. On the other hand, some existing solutions fail to effectively introduce a real-time data and multi-source data fusion mechanism, lacking flexible and efficient data processing capabilities, especially being lagged in response to the rapid fluctuations of new energy output, and it is difficult to adjust the parameters of the protection device, making the setting accuracy of the distribution network protection device insufficient. Summary of the Invention

[0004] Embodiments of the present invention provide a setting method, device, electronic device and storage medium for a distribution network protection device based on new energy output prediction, which can solve the problem of insufficient setting accuracy of the distribution network protection device in the existing technology.

[0005] An embodiment of the present invention provides a setting method for a distribution network protection device based on new energy output prediction, including:

[0006] Obtain the real-time power generation power of new energy units, the start-stop status of new energy units, the environmental data on the power plant side, and the environmental data of the associated weather station; wherein, the environmental data on the power plant side includes the wind speed corresponding to the wind turbine, the wind direction corresponding to the wind turbine, and the light intensity corresponding to the photovoltaic unit; the environmental data of the associated weather station includes the temperature, humidity, wind speed, wind direction and precipitation collected by the associated weather station; the new energy units include wind turbines and photovoltaic units;

[0007] Input the real-time power generation of the wind turbine, the start / stop status of the wind turbine, the wind speed corresponding to the wind turbine, the wind direction corresponding to the wind turbine, and the environmental data of the associated weather station into a preset wind power output prediction model, so that the wind power output prediction generates the output of the wind turbine in a future period;

[0008] Input the real-time power generation of the photovoltaic unit, the start / stop status of the photovoltaic unit, the light intensity corresponding to the photovoltaic unit, and the environmental data of the associated weather station into a preset photovoltaic power output prediction model, so that the photovoltaic power output prediction model generates the output of the photovoltaic unit in a future period;

[0009] Generate the predicted power generation in a future period according to the output of the wind turbine in the future period and the output of the photovoltaic unit in the future period; determine the setting value of the distribution network protection device according to the ratio of the predicted power generation to the theoretical maximum output of the preset new energy unit;

[0010] Set the distribution network protection device according to the setting value.

[0011] Further, the setting value of the distribution network protection device is determined by the following formula:

[0012]

[0013] Where, I n is the setting value of the distribution network protection device; I b is the reference setting value; K u is the setting value increase ratio coefficient; P p is the ratio of the predicted power generation to the theoretical maximum output of the preset new energy unit; K d is the setting value decrease ratio coefficient.

[0014] Further, the training of the wind power output prediction model includes:

[0015] Obtain a wind power output prediction training data set; wherein, the wind power output prediction training data set includes the historical power generation of the wind turbine at the same moment, the start / stop status of the wind turbine, the wind speed corresponding to the wind turbine, the wind direction corresponding to the wind turbine, the environmental data of the associated weather station, and the output label of the wind turbine at a future moment;

[0016] Randomly divide the wind power output prediction training data set into several batches of training samples according to a preset quantity;

[0017] Input the training samples of each batch into the wind power output prediction model in sequence to train the wind power output prediction model until the preset number of training times is reached; wherein, when the wind power output prediction model receives each batch of training samples, it outputs the predicted output of the wind turbine at the future moment corresponding to the training samples; according to the predicted output of the wind turbine at the future moment and the corresponding wind turbine output label at the future moment, calculate the loss function value through the mean square error function; use the Adam optimizer to update the wind power output prediction model according to the loss function value.

[0018] Further, the training of the photovoltaic power output prediction model includes:

[0019] Obtain the photovoltaic power output prediction training data set; wherein, the photovoltaic power output prediction training data set includes the historical power generation of the photovoltaic units at the same moment, the start-stop state of the photovoltaic units, the light intensity corresponding to the photovoltaic units, the environmental data of the associated weather station, and the photovoltaic unit output label at the future moment.

[0020] Randomly divide the photovoltaic power output prediction training data set into several batches of training samples according to the preset quantity.

[0021] Input the training samples of each batch into the photovoltaic power output prediction model in sequence to train the photovoltaic power output prediction model until the preset number of training times is reached; wherein, when the photovoltaic power output prediction model receives each batch of training samples, it outputs the predicted output of the photovoltaic unit at the future moment corresponding to the training samples; according to the predicted output of the photovoltaic unit at the future moment and the corresponding photovoltaic unit output label at the future moment, calculate the loss function value through the mean square error function; use the Adam optimizer to update the photovoltaic power output prediction model according to the loss function value.

[0022] Further, before inputting the training samples of each batch into the wind power output prediction model in sequence to train the wind power output prediction model until the preset number of training times is reached, it further includes:

[0023] For each batch of training samples, perform denoising processing, outlier removal processing, missing value filling processing, and normalization processing to generate updated training samples.

[0024] Further, the associated weather station is the weather station closest to the power plant, and the environmental data on the power plant side is collected in real time through a wind speed sensor deployed at the hub of the wind turbine and a light sensor at the center of the photovoltaic unit array.

[0025] Further, the protection device includes an overcurrent protection device, a low voltage protection device, a high voltage protection device, and a frequency protection device.

[0026] Based on the above method item embodiments, the present invention correspondingly provides device item embodiments.

[0027] An embodiment of the present invention provides a setting device for a distribution network protection device based on new energy output prediction, including: a power plant data acquisition module, a wind turbine output generation module, a photovoltaic unit output generation module, a setting value determination module for the protection device, and a protection device setting module.

[0028] The power plant data acquisition module is used to acquire the real-time power generation power of new energy units, the start-stop status of new energy units, the power plant side environmental data, and the environmental data of the associated meteorological station; wherein, the power plant side environmental data includes the wind speed corresponding to the wind turbine, the wind direction corresponding to the wind turbine, and the light intensity corresponding to the photovoltaic unit; the environmental data of the associated meteorological station includes the temperature, humidity, wind speed, wind direction, and precipitation collected by the associated meteorological station; the new energy units include wind turbines and photovoltaic units;

[0029] The wind turbine output generation module is used to input the real-time power generation power of the wind turbine, the start-stop status of the wind turbine, the wind speed corresponding to the wind turbine, the wind direction corresponding to the wind turbine, and the environmental data of the associated meteorological station into a preset wind power output prediction model, so that the wind power output prediction generates the output of the wind turbine in the future period;

[0030] The photovoltaic unit output generation module is used to input the real-time power generation power of the photovoltaic unit, the start-stop status of the photovoltaic unit, the light intensity corresponding to the photovoltaic unit, and the environmental data of the associated meteorological station into a preset photovoltaic power output prediction model, so that the photovoltaic power output prediction model generates the output of the photovoltaic unit in the future period;

[0031] The setting value determination module for the protection device is used to generate the predicted power generation in the future period according to the output of the wind turbine in the future period and the output of the photovoltaic unit in the future period; and determine the setting value of the distribution network protection device according to the ratio of the predicted power generation to the theoretical maximum output of the preset new energy unit;

[0032] The protection device setting module is used to set the distribution network protection device according to the setting value.

[0033] Based on the above method item embodiments, the present invention correspondingly provides electronic device item embodiments.

[0034] An embodiment of the present invention provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, it can implement the method for setting a distribution network protection device based on new energy output prediction described in any one of the above method item embodiments.

[0035] Based on the above method embodiment, the present invention correspondingly provides a storage medium embodiment.

[0036] An embodiment of the present invention provides a storage medium, on which a computer program is stored. When the computer program is executed by a processor, it can implement the method for setting a distribution network protection device based on new energy output prediction described in any one of the above method embodiments.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] An embodiment of the present invention provides a method, device, electronic device and storage medium for setting a distribution network protection device based on new energy output prediction. The method first collects the real-time power generation, start-stop status, power plant-side environmental data of new energy units and environmental data of associated weather stations. The new energy units include wind turbines and photovoltaic units. Subsequently, the above multi-source information is respectively input into the corresponding wind or photovoltaic output prediction model to generate output prediction values for future periods. Furthermore, the overall predicted power generation is calculated based on the predicted output of wind power and photovoltaic power, and the setting value of the distribution network protection device is determined and set according to the proportion of the predicted power generation in the theoretical maximum output of the new energy unit.

[0039] By introducing the real-time power generation, start-stop status, on-site environmental data of new energy units and multi-dimensional environmental data of associated weather stations, the present invention significantly enhances the fusion processing ability of the output prediction model for multi-source heterogeneous information, and solves the problem of insufficient prediction accuracy caused by relying on a single data source in the prior art. At the same time, by dynamically setting the parameters of the protection device according to the predicted power generation, the adaptive adjustment of the distribution network protection device is realized, effectively coping with the problem of inflexible setting caused by the fluctuation of new energy output, and improving the accuracy of setting the distribution network protection device. Description of the Drawings

[0040] Figure 1 is a schematic flowchart of a method for setting a distribution network protection device based on new energy output prediction provided by an embodiment of the present invention.

[0041] Figure 2 is a schematic structural diagram of a device for setting a distribution network protection device based on new energy output prediction provided by an embodiment of the present invention. Detailed Embodiments

[0042] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0043] As Figure 1 shown, to solve the problem of insufficient accuracy in setting the distribution network protection device in the existing technology, an embodiment of the present invention provides a method for setting the distribution network protection device based on new energy output prediction, which at least includes the following steps:

[0044] Step S1, obtain the real-time power generation power of the new energy unit, the start-stop state of the new energy unit, the environmental data on the power plant side, and the environmental data of the associated weather station;

[0045] In a preferred embodiment, the associated weather station is the weather station closest to the power plant, and the environmental data on the power plant side is collected in real time by a wind speed sensor deployed at the hub of the wind turbine and a light sensor at the center of the photovoltaic array.

[0046] Specifically, the environmental data on the power plant side includes the wind speed corresponding to the wind turbine, the wind direction corresponding to the wind turbine, and the light intensity corresponding to the photovoltaic unit; the environmental data of the associated weather station includes the temperature, humidity, wind speed, wind direction, and precipitation collected by the associated weather station; the new energy unit includes a wind turbine and a photovoltaic unit. By collecting in real time the real-time power generation power, start-stop state, environmental data on the power plant side, and data of the associated weather station of the new energy unit, a multi-dimensional data input system is constructed. Among them, the environmental data on the power plant side is directly related to the equipment operation characteristics (wind speed, wind direction of the wind turbine and light intensity of the photovoltaic unit), and the data of the associated weather station covers regional meteorological indicators such as temperature, humidity, and precipitation. The combination of the two forms a panoramic observation from the microscopic state of the equipment to the macroscopic environmental factors. By synchronously inputting the real-time operation data of the wind and photovoltaic units (such as power generation power fluctuations and start-stop state switches) and environmental parameters (such as wind speed changes at the weather station and the impact of precipitation on light) into a preset wind and photovoltaic output prediction model, the direct fusion of multi-source heterogeneous data is realized, providing input features with high timeliness and strong correlation for output prediction, and effectively overcoming the limitation of traditional methods relying on a single data source. By coupling and correlating equipment-level data (such as power fluctuations of the fan caused by wind speed changes and output responses of the photovoltaic due to changes in light intensity) with environmental parameters (such as the impact of precipitation on the reflectivity of the photovoltaic panel surface and the attenuation effect of temperature on the mechanical efficiency of the fan), a complete data link from the internal working conditions of the equipment to external environmental disturbances is constructed, breaking through the limitation of traditional solutions relying on single-dimensional data (such as only the historical output of the unit or independent data of the weather station), providing a multi-source input basis with high credibility for subsequent analysis, thereby enhancing the analysis ability of data integrity for complex scenarios and ensuring that the subsequent setting calculation can accurately respond to the dynamic coupling relationship between new energy output and environmental factors.

[0047] Step S2: Input the real-time power generation of the wind turbine, the start-stop state of the wind turbine, the wind speed corresponding to the wind turbine, the wind direction corresponding to the wind turbine, and the environmental data of the associated meteorological station into a preset wind power output prediction model, so that the wind power output prediction generates the output of the wind turbine in a future period;

[0048] Specifically, by collecting the operation data of the wind turbine and the associated environmental parameters in real time and inputting them into a preset wind power output prediction model to generate a future output prediction. Specifically, it includes: the current real-time power generation of the wind turbine, the equipment start-stop state, the wind speed and wind direction at the location of the unit, and the temperature, humidity, regional wind speed, wind direction, and precipitation data provided by the associated meteorological station. By integrating the equipment operation status (such as the change in unit efficiency reflected by the power generation power fluctuation and the equipment availability indicated by the start-stop state) and environmental parameters (such as the influence of the wind speed and wind direction monitored by the meteorological station on the force of the unit's blades and the potential interference of precipitation on equipment operation), the wind power output prediction model analyzes the correlation between the current input and historical laws based on the time continuity characteristics of multi-dimensional data, deduces the trend change of wind resources and the unit output response in a future period, and thus outputs the predicted output value of the wind turbine in a future period. This prediction result provides a direct basis for subsequent dynamic adjustment of the setting value, ensuring that the protection device parameters match the real-time changes in wind power output.

[0049] It should be explained here that the wind power output prediction model adopts a Long Short-Term Memory (LSTM) network. LSTM is a variant of the Recurrent Neural Network (RNN) designed specifically for processing time series data. Its core advantage lies in capturing long-term dependencies in the data through a gating mechanism (input gate, forget gate, output gate), while avoiding the vanishing gradient problem of traditional RNNs. In this solution, the input of the LSTM model is multi-dimensional time series data such as the real-time power generation of the wind turbine, start-stop status, wind speed and direction of the unit, and the temperature, humidity, regional wind speed, wind direction, and precipitation of the associated meteorological station. By analyzing the dynamic change laws of the data (such as wind speed trends, instantaneous impacts of start-stop events on output, correction effects of temperature and humidity on air density), the LSTM can effectively model the non-linear mapping relationship between environmental parameters and unit output, and remember key historical patterns (such as output saturation characteristics after continuous strong winds, equipment response delays before and after precipitation). Compared with traditional linear models or static machine learning methods, the core advantages of LSTM in this scenario include: directly processing heterogeneous time series of wind turbine states and meteorological data without manually constructing feature engineering; real-time response to short-term disturbances such as sudden wind speed changes and equipment start-stop, while capturing the long-term impacts of seasonal or periodic meteorological laws on output; automatically filtering out instantaneous noise in sensor data (such as outliers caused by anemometer jitter and communication packet loss) through the forget gate to improve prediction stability. Combining the data input characteristics of this solution, the LSTM model can significantly improve the analysis accuracy of output prediction for complex environmental disturbances, provide highly reliable prediction results for dynamic setting value calculation, and thus reduce the risk of misoperation or refusal to operate of protection devices due to prediction errors.

[0050] In a preferred embodiment, the training of the wind power output prediction model includes:

[0051] Obtain a wind power output prediction training data set; wherein, the wind power output prediction training data set includes historical power generation of wind turbines at the same moment, start-stop status of wind turbines, wind speed corresponding to wind turbines, wind direction corresponding to wind turbines, environmental data of associated meteorological stations, and wind turbine output labels at future moments;

[0052] Randomly divide the wind power output prediction training data set into several batches of training samples according to a preset quantity;

[0053] Input the training samples of each batch into the wind power output prediction model in sequence to train the wind power output prediction model until the preset number of training times is reached; among them, when the wind power output prediction model receives each batch of training samples, it outputs the predicted output of the wind turbine at the future time corresponding to the training samples; according to the predicted output of the wind turbine at the future time and the corresponding wind turbine output label at the future time, calculate the loss function value through the mean square error function; use the Adam optimizer to update the wind power output prediction model according to the loss function value.

[0054] Specifically, the training of the wind power output prediction model is carried out by constructing a highly reliable wind power output prediction training data set and adopting a batch-by-batch iterative training strategy to optimize the model parameters. Specifically, the wind power output prediction training data set is obtained through historical data collection and alignment processing, and contains multi-dimensional input features and corresponding labels at the same timestamp: the input features include the historical power generation power of the wind turbine (reflecting the historical output capacity of the equipment), the start-stop state (identifying the operation / stop state of the equipment), the wind speed and direction at the location of the unit (the core environmental parameters directly driving the power generation), and the environmental data such as temperature and humidity of the associated weather station (indirectly affecting the unit efficiency and grid operation conditions); the label is the actual output value of the wind turbine at the future time (such as the next 15 minutes), which is used to supervise the deviation calculation between the model output and the true value. When constructing the data set, data cleaning and alignment are required: perform denoising processing on the original data (such as filtering abnormal jump values of sensors), filling missing values (using linear interpolation or mean filling of adjacent time periods), and strictly aligning multi-source data according to the timestamp to ensure the spatio-temporal consistency of the input features and labels. After the preprocessing is completed, the data set is randomly divided into several batches according to the preset batch size (such as 64 or 128 samples / batch) to improve the model training efficiency and avoid local overfitting. During the training process, each batch of data is input into the wind power output prediction model in sequence. The model outputs the predicted future output value based on the multi-dimensional features of the current batch (historical power, wind speed, meteorological data, etc.). The deviation degree between the predicted value and the true label is quantified through the mean square error function (MSE) to generate the loss function value. Utilize the adaptive learning rate characteristic of the Adam optimizer to update the model weights according to the loss function gradient: the initial learning rate is set to 0.001, and the cosine annealing strategy is adopted for dynamic adjustment to balance the convergence speed and accuracy; at the same time, the early stopping method (Early Stopping) is introduced to terminate the training when the validation set loss does not decrease for 10 consecutive training cycles to prevent overfitting. Through multiple rounds of iterative optimization, the model gradually learns the non-linear relationship between wind speed and power, the instantaneous impact of start-stop events on the output, and the coupling effect of environmental parameters, and finally obtains a prediction model that can accurately predict the output fluctuation of the wind turbine, providing a highly reliable input for the calculation of dynamic setting values.

[0055] In a preferred embodiment, before the training samples of each batch are sequentially input into the wind power output prediction model and the wind power output prediction model is trained until the preset number of training times is reached, the following steps are further included:

[0056] For each batch of training samples, denoising processing, outlier removal processing, missing value filling processing, and normalization processing are performed to generate updated training samples.

[0057] It should be noted here that denoising processing refers to removing unnecessary noise or interference signals from the data. Common denoising methods include low-pass filters (used to remove high-frequency noise), Kalman filters (used for state estimation of dynamic systems), and wavelet transforms, etc.; outliers refer to those data points that significantly deviate from the normal range. Common outlier detection methods include machine learning-based methods (such as isolation forest, DBSCAN clustering algorithm); missing values refer to the situation where some data points in the dataset are not recorded. Common filling methods include interpolation method, median filling, nearest neighbor filling, and output prediction model filling, etc.; normalization is to convert data of different magnitudes to the same scale, usually between [0, 1].

[0058] In one embodiment, the present invention divides the wind power output prediction training dataset into a training set, a validation set, and a test set according to the ratio of 6:2:2, and constructs a model performance evaluation system: the training set is used for model parameter optimization, and the weights and biases are adjusted by minimizing the loss function (such as mean square error); the validation set is used for hyperparameter tuning (including learning rate, batch size, number of network layers, and number of neurons), and based on root mean square error (RMSE), mean absolute error (MAE), coefficient of determination (R 2 ) and other indicators to screen the optimal model configuration and calculate the comprehensive performance indicators; the test set finally evaluates the model generalization ability. In addition to regression indicators, the accuracy, precision, and recall rate in the classification scenario are combined to comprehensively verify the prediction reliability of the model on unknown data to ensure that it meets the actual engineering application requirements.

[0059] Specifically, in one embodiment, the real-time operation data is combined with the new energy output prediction result according to the timestamp, and the comprehensive performance indicator is calculated. The expression is:

[0060]

[0061] where C is the comprehensive performance indicator; S is the mean square error; A is the mean absolute error; R 2 is the coefficient of determination.

[0062] It should be noted that S and A reflect the magnitude of the prediction error of the output prediction model. After standardization, they can be compared on the same scale. The smaller the value, the smaller the error. R 2It reflects the ability of the output prediction model to explain data variability. The larger the value, the better the fitting effect of the output prediction model. The closer the C value is to 1, the better the overall performance of the output prediction model. The closer the C value is to 0, the worse the performance of the output prediction model.

[0063] In addition, in one embodiment, based on historical new energy output data, power grid operation data, and weather condition data, an expected target is defined, and the expression is:

[0064]

[0065] where F is the expected target;

[0066] Compare the comprehensive performance index C with the expected target F to obtain the gap index F, where D = max(0, F - C). When C ≥ F, D = 0, indicating that the performance of the output prediction model has reached or exceeded the expected target and no further optimization is required. When C < F, D > 0, indicating that the performance of the output prediction model has not reached the expected target, and the larger the gap, the more optimization is needed.

[0067] Step S3: Input the real-time power generation power of the photovoltaic unit, the start-stop state of the photovoltaic unit, the corresponding light intensity of the photovoltaic unit, and the environmental data of the associated meteorological station into a preset photovoltaic output prediction model, so that the photovoltaic output prediction model generates the output of the photovoltaic unit in the future time period;

[0068] Specifically, the present invention collects the operation data of the photovoltaic unit and the associated environmental parameters in real time and inputs them into a preset photovoltaic output prediction model to generate a future output prediction. Specifically, it includes: the current real-time power generation power of the photovoltaic unit, the equipment start-stop state, the light intensity in the area where the photovoltaic array is located, and the temperature, humidity, and precipitation data provided by the associated meteorological station. By integrating the equipment operation status (such as the change in component efficiency reflected by the power generation power fluctuation and the equipment availability indicated by the start-stop state) and environmental parameters (such as the direct impact of light intensity on component output, the attenuation effect of temperature on the performance of semiconductor materials, and the cleaning or blocking effect of precipitation on the panel surface), based on the time-series correlation of multi-dimensional data, the photovoltaic output prediction model analyzes the mapping relationship between the current input and the historical output law, deduces the change trend of light resources in the future time period and the comprehensive effect of environmental conditions on the output, and thus outputs the predicted output value of the photovoltaic unit in the future time period. This prediction result provides a direct basis for subsequent dynamic adjustment of the setting value, ensures that the protection device parameters match the real-time fluctuation of the photovoltaic output, and reduces the risk of protection malfunction or refusal due to prediction deviation.

[0069] In this technical solution, the photovoltaic output prediction model adopts a Long Short-Term Memory (LSTM) network. LSTM is a recurrent neural network designed for time series data. It captures temporal dependencies in data through a gating mechanism (input gate, forget gate, output gate), and is particularly good at modeling complex sequences with long-term periodicity and short-term mutation characteristics. In this scenario, the inputs to the LSTM model are the real-time power generation of the photovoltaic unit, its start-stop status, light intensity, and temperature, humidity, and precipitation data of the associated weather station. These parameters form a multi-dimensional time series. For example, the dynamic relationship between light intensity and power generation, the non-linear effect of temperature on component efficiency, and the triggering effect of precipitation events on output mutations.

[0070] In a preferred embodiment, the training of the photovoltaic output prediction model includes:

[0071] Obtain a photovoltaic output prediction training data set; wherein, the photovoltaic output prediction training data set includes historical power generation of the photovoltaic unit, the start-stop status of the photovoltaic unit, the light intensity corresponding to the photovoltaic unit, environmental data of the associated weather station, and the output label of the photovoltaic unit at a future time at several same moments.

[0072] Randomly divide the photovoltaic output prediction training data set into several batches of training samples according to a preset quantity.

[0073] Input each batch of training samples into the photovoltaic output prediction model in sequence to train the photovoltaic output prediction model until a preset number of training times is reached; wherein, when the photovoltaic output prediction model receives each batch of training samples, it outputs the predicted output of the photovoltaic unit at the future time corresponding to the training samples; according to the predicted output of the photovoltaic unit at the future time and the corresponding output label of the photovoltaic unit at the future time, calculate the loss function value through the mean square error function; use the Adam optimizer to update the photovoltaic output prediction model according to the loss function value.

[0074] In a preferred embodiment, before inputting each batch of training samples into the photovoltaic output prediction model in sequence to train the photovoltaic output prediction model until a preset number of training times is reached, it further includes:

[0075] For each batch of training samples, perform denoising processing, outlier removal processing, missing value filling processing, and normalization processing to generate updated training samples.

[0076] Step S4: Generate the predicted power generation for the future period based on the output of the wind turbine and the output of the photovoltaic unit in the future period; determine the setting value of the distribution network protection device according to the ratio of the predicted power generation to the theoretical maximum output of the preset new energy unit.

[0077] Specifically, by integrating the output prediction results of wind power and photovoltaic units, a dynamic evaluation and setting value adjustment system for new energy power generation is constructed. Specifically, the output of wind turbines in the future period generated by the wind power output prediction model is superimposed with the output of photovoltaic units generated by the photovoltaic output prediction model to calculate the predicted total power generation in the future period. This total power generation reflects the comprehensive output capacity of new energy units under specific environmental conditions and equipment states. Further, the predicted total power generation is divided by the total output value of new energy units under the theoretical maximum operating state (such as the sum of the rated power of wind turbines and the nominal power of photovoltaic modules) to obtain the predicted power generation ratio parameter.

[0078] In a preferred embodiment, the setting value of the distribution network protection device is determined by the following formula:

[0079]

[0080] where, I n is the setting value of the distribution network protection device; I b is the reference setting value; K u is the setting value increase ratio coefficient; P p is the ratio of the predicted power generation to the theoretical maximum output of the preset new energy unit; K d is the setting value decrease ratio coefficient.

[0081] Specifically, the ratio of the predicted power generation to the theoretical maximum output of the preset new energy unit quantifies the deviation degree between the new energy output and the theoretical limit. For example, when the predicted power generation ratio reaches 80%, it indicates that the new energy output is close to the peak value, and the power grid may face risks such as increased short-circuit current and voltage fluctuation; while when the ratio is lower than 30%, it indicates that the output is seriously insufficient, and there is a hidden danger that the fault current is lower than the traditional setting threshold. Based on this ratio parameter, a segmented setting rule is adopted to dynamically adjust the action threshold of the protection device: when the ratio exceeds the preset upper limit, the setting value is adjusted upward in proportion to avoid misoperation; when the ratio is lower than the preset lower limit, the setting value is adjusted downward correspondingly to prevent refusal to operate; and the reference setting value is maintained within the normal range.

[0082] It should be noted here that the reference setting value is the standard setting value when there is no new energy access, which is the initial threshold set based on the protection requirements of the traditional distribution network. In one embodiment, the setting value increase ratio coefficient K u ∈[0.05, 0.15], and the setting value decrease ratio coefficient K d ∈[0.03, 0.10].

[0083] Step S5: Set the distribution network protection device according to the setting value.

[0084] Specifically, after determining the setting value by predicting the power generation ratio, the calculation result is directly applied to the parameter adjustment of the distribution network protection device. Specifically, according to the ratio of the predicted power generation to the theoretical maximum output, the action thresholds such as overcurrent and low voltage of the protection device are numerically corrected according to the preset rules. For example, when the predicted power generation ratio is high, the action current threshold of the overcurrent protection is correspondingly increased to adapt to the increase in the short-circuit current level caused by the sudden increase in new energy output; when the ratio is low, the trigger threshold of the low voltage protection is reduced to avoid delayed action due to the voltage drop not reaching the original threshold when the output is insufficient. After the setting value is adjusted, the protection device immediately loads the new parameters to make its judgment logic match the current new energy output state in real time. By periodically updating the setting value (such as adjusting it based on the latest prediction result every 15 minutes), it is ensured that the protection device always dynamically optimizes following the fluctuations of new energy output, so as to quickly and accurately act when a fault occurs, preventing both mis-tripping of normal lines and missed detection of real faults, and improving the reliability and safety of power grid operation.

[0085] In a preferred embodiment, the protection device includes an overcurrent protection device, a low voltage protection device, a high voltage protection device, and a frequency protection device.

[0086] Specifically, each device is provided with an independent reference setting value and is differentially adjusted according to the new energy output prediction result. Specifically:

[0087] The reference setting value I of the overcurrent protection device b_oc is set according to the rated current of the line, and the dynamic adjustment formula is:

[0088]

[0089] where I n_oc is the setting value of the overcurrent protection device; V b_oc is the reference setting value of the overcurrent protection device;

[0090] The reference setting value V of the low voltage protection device b_lv is set based on the safety lower limit of the bus voltage, and the adjustment formula is:

[0091]

[0092] where V n_lv is the setting value of the low voltage protection device; V b_lv is the reference setting value of the low voltage protection device;

[0093] The reference setting value V of the high voltage protection device b_hv corresponds to the voltage safety upper limit, and the adjustment formula is:

[0094]

[0095] Among them, V n_hv is the setting value of the high-voltage protection device; V b_hv is the reference setting value of the high-voltage protection device;

[0096] The reference setting value f of the frequency protection device b Based on the allowable deviation range of the system frequency, the adjustment formula is:

[0097]

[0098] Among them, f n is the setting value of the frequency protection device; f b is the reference setting value of the frequency protection device.

[0099] Through the above differential adjustment, the setting values of each protection device are accurately matched with the new energy output state respectively, ensuring the action accuracy and coordination of overcurrent, high and low voltage and frequency protection under different working conditions.

[0100] In a preferred embodiment, before setting the distribution network protection device according to the setting value, a simulation model can be built to verify the effect of the setting. Build a simulation model including new energy power generation equipment, power grid and distribution network protection device, input the adjusted parameters of the distribution network protection device and the prediction results of new energy output into the simulation model, and verify the effect of the distribution network protection device;

[0101] It should be noted that a power simulation model is built through the MATLAB interactive environment and Simulink design tool. First, input historical meteorological data and new energy output prediction data into the MATLAB interactive environment, and then create sub-units in Simulink to represent photovoltaic panels, wind turbines, power grid and distribution network protection devices respectively, and configure the parameters of each sub-unit, such as rated power, efficiency curve, etc. Connect each sub-unit through the interface to form a complete power simulation model. Input the adjusted parameters of the distribution network protection device and the prediction results of new energy output into the power simulation model, set the simulation time and step size, run the simulation to obtain the results, use the analysis tools of MATLAB to analyze the simulation results, generate a report. In the power simulation model, simulate the power generation characteristics of new energy power generation equipment, simulate the structure of the power grid, load distribution, voltage level and frequency stability and other indicators, simulate the overcurrent protection, low voltage protection and high voltage protection of the distribution network protection device, etc.; create multiple scenarios to simulate different situations that may be encountered in actual operation, including significant increase or decrease of new energy output, load fluctuation and power generation fluctuation caused by weather change, record the action of the distribution network protection device under each working condition, including whether the protection mechanism is triggered, the triggering time point and the impact on the power grid operation, etc. Through these action situations, the accuracy of the prediction results of new energy output can be verified;

[0102] Based on the above method embodiment, the present invention correspondingly provides an apparatus embodiment.

[0103] As Figure 2 shown, an embodiment of the present invention provides a setting device for a distribution network protection device based on new energy output prediction, including: a power plant data acquisition module, a wind turbine output generation module, a photovoltaic unit output generation module, a setting value determination module for the protection device, and a protection device setting module;

[0104] The power plant data acquisition module is configured to acquire the real-time power generation power of new energy units, the start-stop state of new energy units, the power plant side environmental data, and the environmental data of the associated weather station; wherein, the power plant side environmental data includes the wind speed corresponding to the wind turbine, the wind direction corresponding to the wind turbine, and the light intensity corresponding to the photovoltaic unit; the environmental data of the associated weather station includes the temperature, humidity, wind speed, wind direction, and precipitation collected by the associated weather station; the new energy units include wind turbines and photovoltaic units;

[0105] The wind turbine output generation module is configured to input the real-time power generation power of the wind turbine, the start-stop state of the wind turbine, the wind speed corresponding to the wind turbine, the wind direction corresponding to the wind turbine, and the environmental data of the associated weather station into a preset wind power output prediction model, so that the wind power output prediction generates the output of the wind turbine in a future period;

[0106] The photovoltaic unit output generation module is configured to input the real-time power generation power of the photovoltaic unit, the start-stop state of the photovoltaic unit, the light intensity corresponding to the photovoltaic unit, and the environmental data of the associated weather station into a preset photovoltaic power output prediction model, so that the photovoltaic power output prediction model generates the output of the photovoltaic unit in a future period;

[0107] The setting value determination module for the protection device is configured to generate the predicted power generation in a future period according to the output of the wind turbine in the future period and the output of the photovoltaic unit in the future period; and determine the setting value of the distribution network protection device according to the ratio of the predicted power generation to the theoretical maximum output of the preset new energy unit;

[0108] The protection device setting module is configured to set the distribution network protection device according to the setting value.

[0109] It should be noted that the embodiments of the device described above correspond to the above embodiments of the present invention, and can implement the method for setting a distribution network protection device based on new energy output prediction described in any one of the above of the present invention. In addition, the embodiments of the above device are merely illustrative. The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. In addition, in the accompanying drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that they have a communication connection, which can be specifically implemented as one or more communication buses or signal lines. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0110] Based on the above method embodiment of the present invention, a corresponding electronic device embodiment is provided.

[0111] An embodiment of the present invention provides an electronic device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, the method for setting a distribution network protection device based on new energy output prediction described in any one of the present invention is implemented, or when the processor executes the computer program, the functions of each module in the above device embodiments are implemented.

[0112] Exemplarily, the computer program can be divided into one or more modules. The one or more modules are stored in the memory and executed by the processor to complete the present invention. The one or more modules can be a series of computer program instruction segments capable of completing specific functions, and the instruction segments are used to describe the execution process of the computer program in the terminal device.

[0113] The terminal device can be a computing device such as a desktop computer, a notebook, a palm computer, and a cloud server. The terminal device may include, but is not limited to, a processor and a memory.

[0114] The so-called processor may be a Central Processing Unit (CPU), or may also be other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field-Programmable Gate Arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The processor is the control center of the terminal device and connects all parts of the entire terminal device through various interfaces and circuits.

[0115] The memory can be used to store the computer programs and / or modules. The processor realizes various functions of the terminal device by running or executing the computer programs and / or modules stored in the memory, and by calling the data stored in the memory. The memory mainly includes a program storage area and a data storage area. Among them, the program storage area can store an operating system, application programs required for at least one function, etc.; the data storage area can store data created according to the use of the mobile phone, etc. In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, at least one magnetic disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0116] Based on the above method item embodiments, the present invention correspondingly provides storage medium item embodiments;

[0117] Another embodiment of the present invention provides a storage medium. The storage medium includes a stored computer program. When the computer program runs, it controls the device where the storage medium is located to execute any one of the above-mentioned distribution network protection device setting methods based on new energy output prediction of the present invention.

[0118] Among them, the above storage medium is a computer-readable storage medium, and the computer program includes computer program code, which can be in the form of source code, object code, executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disc, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc. It should be noted that the content included in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electrical carrier signals and telecommunication signals.

[0119] In the description of this specification, the description referring to terms such as "one embodiment", "some embodiments", "example", "specific example" or "some examples", etc. means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. Moreover, the specific features, structures, materials or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0120] The above is the preferred embodiment of the present invention. It should be pointed out that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can still be made, and these improvements and refinements are also regarded as the protection scope of the present invention.

Claims

1. A setting method for a distribution network protection device based on new energy output prediction, characterized in that, Including: Obtain the real-time power generation power of the new energy unit, the start-stop state of the new energy unit, the environmental data on the power plant side, and the environmental data of the associated meteorological station; wherein, the environmental data on the power plant side includes the wind speed corresponding to the wind turbine, the wind direction corresponding to the wind turbine, and the light intensity corresponding to the photovoltaic unit; the environmental data of the associated meteorological station includes the temperature, humidity, wind speed, wind direction, and precipitation collected by the associated meteorological station; the new energy unit includes a wind turbine and a photovoltaic unit; Input the real-time power generation power of the wind turbine, the start-stop state of the wind turbine, the wind speed corresponding to the wind turbine, the wind direction corresponding to the wind turbine, and the environmental data of the associated meteorological station into a preset wind power output prediction model, so that the wind power output prediction generates the power output of the wind turbine in the future period; Input the real-time power generation power of the photovoltaic unit, the start-stop state of the photovoltaic unit, the light intensity corresponding to the photovoltaic unit, and the environmental data of the associated meteorological station into a preset photovoltaic power output prediction model, so that the photovoltaic power output prediction model generates the power output of the photovoltaic unit in the future period; Generate the predicted power generation in the future period according to the power output of the wind turbine in the future period and the power output of the photovoltaic unit in the future period; determine the setting value of the distribution network protection device according to the ratio of the predicted power generation to the theoretical maximum output of the preset new energy unit; Perform setting on the distribution network protection device according to the setting value.

2. The setting method of the distribution network protection device based on new energy output prediction according to claim 1, characterized in that, Determine the setting value of the distribution network protection device through the following formula: Among them, I n is the setting value of the distribution network protection device; I b is the reference setting value; K u is the proportional coefficient for increasing the setting value; P p is the ratio of the predicted power generation to the theoretical maximum output of the preset new energy unit; K d is the proportional coefficient for decreasing the setting value.

3. The setting method of the distribution network protection device based on new energy output prediction according to claim 2, characterized in that, The training of the wind power output prediction model includes: Obtain a wind power output prediction training data set; wherein, the wind power output prediction training data set includes the historical power generation power of the wind turbine, the start-stop state of the wind turbine, the wind speed corresponding to the wind turbine, the wind direction corresponding to the wind turbine, the environmental data of the associated meteorological station, and the wind turbine output label at a future time at several same moments; Randomly divide the wind power output prediction training data set into several batches of training samples according to a preset quantity; Input each batch of training samples into the wind power output prediction model in sequence to train the wind power output prediction model until a preset number of training times is reached; wherein, when the wind power output prediction model receives each batch of training samples, it outputs the predicted wind turbine output at the future time corresponding to the training samples; according to the predicted wind turbine output at the future time and the corresponding wind turbine output label at the future time, calculate the loss function value through the mean square error function; use the Adam optimizer to update the wind power output prediction model according to the loss function value.

4. The setting method of the distribution network protection device based on new energy output prediction according to claim 3, characterized in that, The training of the photovoltaic power output prediction model includes: Obtain a photovoltaic power output prediction training data set; wherein, the photovoltaic power output prediction training data set includes the historical power generation power of the photovoltaic unit, the start-stop state of the photovoltaic unit, the light intensity corresponding to the photovoltaic unit, the environmental data of the associated meteorological station, and the photovoltaic unit output label at a future time at several same moments; Randomly divide the photovoltaic power output prediction training data set into several batches of training samples according to a preset quantity; Input the training samples of each batch into the photovoltaic output prediction model in sequence to train the photovoltaic output prediction model until the preset number of training times is reached; among them, when the photovoltaic output prediction model receives each batch of training samples, it outputs the predicted output of the photovoltaic unit at the future moment corresponding to the training samples; according to the predicted output of the photovoltaic unit at the future moment and the corresponding photovoltaic unit output label at the future moment, calculate the loss function value through the mean square error function; use the Adam optimizer to update the photovoltaic output prediction model according to the loss function value.

5. The setting method of the distribution network protection device based on new energy output prediction according to claim 4, characterized in that, Before the step of inputting the training samples of each batch into the wind power output prediction model in sequence to train the wind power output prediction model until the preset number of training times is reached, it further includes: For each batch of training samples, perform denoising processing, outlier removal processing, missing value filling processing, and normalization processing to generate updated training samples.

6. The setting method of the distribution network protection device based on new energy output prediction according to claim 5, wherein, The associated weather station is the weather station closest to the power plant, and the environmental data on the power plant side is collected in real time through a wind speed sensor deployed at the hub of the wind turbine and a light sensor at the center of the photovoltaic array.

7. The setting method of the distribution network protection device based on new energy output prediction according to claim 6, characterized in that, The protection device includes an overcurrent protection device, a low voltage protection device, a high voltage protection device, and a frequency protection device.

8. A setting device for a distribution network protection device based on new energy output prediction, characterized in that, It includes: A power plant data acquisition module, a wind turbine output generation module, a photovoltaic unit output generation module, a setting value determination module for the protection device, and a protection device setting module; The power plant data acquisition module is used to acquire the real-time power generation power of the new energy units, the start-stop states of the new energy units, the environmental data on the power plant side, and the environmental data of the associated weather station; among them, the environmental data on the power plant side includes the wind speed corresponding to the wind turbine, the wind direction corresponding to the wind turbine, and the light intensity corresponding to the photovoltaic unit; the environmental data of the associated weather station includes the temperature, humidity, wind speed, wind direction, and precipitation collected by the associated weather station; the new energy units include wind turbines and photovoltaic units; The wind turbine output generation module is used to input the real-time power generation power of the wind turbine, the start-stop state of the wind turbine, the wind speed corresponding to the wind turbine, the wind direction corresponding to the wind turbine, and the environmental data of the associated weather station into a preset wind power output prediction model, so that the wind power output prediction model generates the output of the wind turbine in the future period; The photovoltaic unit output generation module is used to input the real-time power generation power of the photovoltaic unit, the start-stop state of the photovoltaic unit, the light intensity corresponding to the photovoltaic unit, and the environmental data of the associated weather station into a preset photovoltaic output prediction model, so that the photovoltaic output prediction model generates the output of the photovoltaic unit in the future period; The setting value determination module for the protection device is used to generate the predicted power generation in the future period according to the output of the wind turbine in the future period and the output of the photovoltaic unit in the future period; determine the setting value of the distribution network protection device according to the ratio of the predicted power generation to the theoretical maximum output of the preset new energy units; The protection device setting module is used to set the distribution network protection device according to the setting value.

9. An electronic device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the setting method of the distribution network protection device based on new energy output prediction according to any one of claims 1 to 7.

10. A storage medium, characterized in that, The storage medium includes a stored computer program. When the computer program runs, it controls the device where the storage medium is located to execute the setting method of the distribution network protection device based on new energy output prediction according to any one of claims 1 to 7.

Images