Power supply recovery method and device of power distribution network, electronic equipment and storage medium

Abstract

The invention discloses a power supply recovery method and device of a power distribution network, electronic equipment and a storage medium. The method comprises the steps that historical output data, historical weather data and future weather data of a distributed power supply in a power distribution network are acquired, and the distributed power supply at least comprises a photovoltaic power generation assembly and a wind power generation assembly; utilizing the pre-trained neural network model to predict the future output of the distributed power supply based on the historical output data, the historical weather data and the future weather data to obtain a future output prediction result; and based on the future output prediction result, performing load shedding operation on the distributed power supply to obtain a power supply recovery result, the power supply recovery result being used for representing a power supply recovery result obtained after the load shedding operation is performed on the distributed power supply. According to the invention, the technical problem of low power restoration efficiency of the power distribution network in the related art is solved.

Description

Technical Field

[0001] The present invention relates to the field of power supply restoration, and in particular to a method, device, electronic equipment and storage medium for power supply restoration of a distribution network. Background Art

[0002] The distribution network refers to the power grid that receives energy from the power supply side and distributes it to various types of power users locally or remotely through distribution equipment. It plays a relatively important role in the entire power transmission process. When equipment failure, line failure, overvoltage or undervoltage failure, frequency failure and other accidents occur in the distribution network, they may cause power outages, resulting in equipment loss, property loss and even threats to personal safety. The problem of power supply recovery after power outages caused by distribution network accidents has become one of the main technical issues studied in the field of distribution network power supply technology.

[0003] At present, in the related technologies, the power supply restoration for distribution network is carried out, or a distribution network fault restoration strategy for maximizing the restoration benefit is proposed, and the restoration benefit, restoration cost and restoration risk of the power outage node are taken as the objective function and the objective function is optimized and solved. However, this method requires the mobile emergency power supply to cooperate with the power supply partition optimization. When the mobile emergency power supply is difficult to dispatch, the output is insufficient or even the mobile emergency power supply fails, the restoration strategy is difficult to implement. Or a wind-solar-storage joint output model is proposed and the model is solved using an improved particle swarm algorithm. However, the algorithm requires the reconstruction of part of the power grid and the main grid, and the input data volume is large and the computing resource occupancy rate is high. At the same time, it is not clear how to calculate the maximum output value of wind storage or solar storage. Or a distribution network post-disaster power supply restoration model for user power outage loss assessment is proposed, and the user power outage loss is introduced in the construction of the objective function. However, the model still faces the same problem as the existing distribution network restoration model solution, that is, the problem of linearizing the nonlinear part in the power flow constraint. If the nonlinear part in the power flow constraint is approximated, it may cause power restoration delay, which is difficult to meet the demand for rapid power restoration. In summary, the efficiency of power supply restoration for distribution network in the related technologies is low.

[0004] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0005] The embodiments of the present invention provide a power supply restoration method, device, electronic device and storage medium for a distribution network, so as to at least solve the technical problem of low efficiency in power supply restoration for a distribution network in the related art.

[0006] According to one aspect of an embodiment of the present invention, a power supply restoration method for a distribution network is provided, comprising: obtaining historical output data, historical weather data and future weather data of distributed power sources in the distribution network, wherein the distributed power sources include at least photovoltaic power generation components and wind power generation components; using a pre-trained neural network model, based on the historical output data, historical weather data and future weather data, predicting the future output of the distributed power sources to obtain a future output prediction result, wherein the pre-trained neural network model is used to represent a pre-trained neural network model for predicting the future output of the distributed power sources; based on the future output prediction result, performing a load shedding operation on the distributed power sources to obtain a power supply restoration result, wherein the power supply restoration result is used to represent a power supply restoration result obtained after the load shedding operation is performed on the distributed power sources.

[0007] Optionally, a pre-trained neural network model is used to predict the future output of distributed power sources based on historical output data, historical weather data and future weather data to obtain future output prediction results, including: determining the degree of overlap between future weather and historical weather of distributed power sources based on historical weather data and future weather data; and obtaining future output prediction results based on the degree of overlap, historical output data, historical weather data and future weather data using a pre-trained neural network model.

[0008] Optionally, a pre-trained neural network model is used to obtain future output prediction results based on the overlap, historical output data, historical weather data and future weather data, including: determining the future weather prediction value of the distributed power source based on the overlap and preset conditions, wherein the preset conditions are used to represent pre-set judgment conditions for measuring the overlap; and obtaining future output prediction results based on the overlap, preset conditions, historical output data, historical weather data and future weather prediction values ​​using a pre-trained neural network model.

[0009] Optionally, a pre-trained neural network model is used to obtain future output prediction results based on the degree of overlap, preset conditions, historical output data, historical weather data and future weather forecast values, including: based on the degree of overlap and preset conditions, historical output data and historical weather data are screened to obtain target historical output data and target historical weather data; and the target historical output data, target historical weather data and future weather forecast values ​​are used as input variables of the pre-trained neural network model to obtain future output prediction results.

[0010] Optionally, the method further includes: obtaining an initial neural network model; and training the initial neural network model based on future weather forecast values ​​to obtain a pre-trained neural network model.

[0011] Optionally, based on historical weather data and future weather data, the degree of overlap between future weather and historical weather of distributed power sources is determined, including: determining the amount of change in historical weather data based on historical data at different times in the historical weather data; constructing the amount of change in future weather data based on a target number of samples and future weather data, wherein the target number of samples is used to represent the number of samples of future weather data; determining the degree of overlap based on the amount of change in historical weather data and the amount of change in future weather data.

[0012] According to another aspect of an embodiment of the present invention, there is also provided a power supply recovery device for a distribution network, comprising: an acquisition module, used to obtain historical output data, historical weather data and future weather data of distributed power sources in the distribution network, wherein the distributed power sources include at least photovoltaic power generation components and wind power generation components; a prediction module, used to use a pre-trained neural network model to predict the future output of the distributed power sources based on the historical output data, historical weather data and future weather data, and obtain a future output prediction result, wherein the pre-trained neural network model is used to represent a pre-trained neural network model for predicting the future output of the distributed power sources; an execution module, used to perform a load shedding operation on the distributed power sources based on the future output prediction result, and obtain a power supply recovery result, wherein the power supply recovery result is used to represent a power supply recovery result obtained after the load shedding operation is performed on the distributed power sources.

[0013] According to another aspect of an embodiment of the present invention, there is further provided an electronic device, comprising: a memory storing an executable program; and a processor for running the program, wherein the method in each embodiment of the present invention is executed when the program is running.

[0014] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is provided. The computer-readable storage medium includes a stored executable program, wherein when the executable program is running, the device where the computer-readable storage medium is located is controlled to execute the methods in various embodiments of the present invention.

[0015] According to another aspect of an embodiment of the present invention, a computer program product is provided, including a computer program. When the computer program is executed by a processor, the method in each embodiment of the present invention is implemented.

[0016] According to another aspect of an embodiment of the present invention, a computer program product is provided, including a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method in each embodiment of the present invention is implemented.

[0017] According to another aspect of the embodiments of the present invention, a computer program is further provided. When the computer program is executed by a processor, the methods in the embodiments of the present invention are implemented.

[0018] In an embodiment of the present invention, a power supply restoration method for a distribution network is provided, comprising: obtaining historical output data, historical weather data and future weather data of distributed power sources in the distribution network, wherein the distributed power sources include at least photovoltaic power generation components and wind power generation components; using a pre-trained neural network model, based on the historical output data, historical weather data and future weather data, predicting the future output of the distributed power sources to obtain a future output prediction result, wherein the pre-trained neural network model is used to represent a pre-trained neural network model for predicting the future output of the distributed power sources; based on the future output prediction result, performing a load shedding operation on the distributed power sources to obtain a power supply restoration result, wherein the power supply restoration result is used to represent a power supply restoration result obtained after the load shedding operation is performed on the distributed power sources. It is easy to notice that the present application can predict future output in a relatively short time by acquiring historical output data, historical weather data and future weather data, and using a pre-trained neural network model to predict future output, thereby quickly formulating load shedding operation strategies and improving the efficiency of power supply restoration. By more accurately predicting future output conditions, excessive or insufficient power supply restoration operations can be avoided, thereby reducing the cost of power supply restoration. At the same time, it can better cope with the volatility and uncertainty of distributed power sources in the distribution network, improve the reliability and stability of the power supply system, and thus solve the technical problem of low efficiency of power supply restoration for distribution networks in related technologies. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0020] Figure 1 is a flow chart of a power supply restoration method for a distribution network according to an embodiment of the present invention;

[0021] Figure 2 is a flow chart of an optional power supply restoration method of a distribution network according to an embodiment of the present invention;

[0022] Figure 3 is a schematic diagram of a power supply restoration device for a distribution network according to an embodiment of the present invention. DETAILED DESCRIPTION

[0023] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0024] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0025] According to one aspect of an embodiment of the present invention, a power supply restoration method for a distribution network is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in an order different from that shown here.

[0026] Figure 1 is a flow chart of a power supply restoration method for a distribution network according to an embodiment of the present invention. Figure 1 As shown, the method comprises the following steps:

[0027] Step S102, obtaining historical output data, historical weather data and future weather data of distributed power sources in the distribution network.

[0028] Among them, the distributed power source includes at least photovoltaic power generation components and wind power generation components.

[0029] The above-mentioned distributed power source may refer to a small independent power generation device distributed in the distribution network. The distributed power source may be a photovoltaic power generation component, a wind power generation component, a micro gas generator, etc. Distributed power sources are usually closer to the power load than traditional centralized power plants, and can respond to demand changes more quickly and improve power supply reliability. Photovoltaic power generation components may be devices that use the solar photovoltaic effect to convert solar energy into electrical energy. Photovoltaic power generation systems may consist of photovoltaic components, inverters, brackets, and wiring, etc., and may be flexibly installed on the roof, ground, or other open space of a building. Photovoltaic power generation systems may be expanded according to demand to achieve distributed power generation. Wind power generation components may use wind power to drive wind turbines to rotate and generate electricity. Wind power generation systems may consist of wind turbines, converters, control systems, etc., and may be installed in high-rise buildings, mountaintops, plains, etc. Wind power generation systems may be laid out according to wind resources in different regions to achieve distributed power generation of renewable energy. The above-mentioned distributed power sources may also be determined according to actual needs, and are not limited here.

[0030] The above historical output data can refer to the power output of distributed power sources over a period of time in the past, which can be recorded hourly or daily. These data can help analyze the operation, stability and reliability of distributed power sources and provide a reference for power supply restoration.

[0031] The above-mentioned historical weather data may refer to the meteorological conditions in the area where the distributed power source is located in the past period of time, including information such as temperature, humidity, wind speed, precipitation, etc. These data may affect the power generation capacity and output of the distributed power source.

[0032] The above-mentioned future weather data may refer to the forecast data of the meteorological conditions in the area where the distributed power source is located in the future period of time. It can be obtained through weather forecasts and other means, and may include information such as weather conditions, temperature changes, wind speed and direction, etc. It can help predict the future power output of the distributed power source and provide a reference for power supply restoration and grid dispatching.

[0033] In an optional embodiment, the historical output data of the distributed power source can be obtained, and the actual power generation data of the distributed power source in the past period of time can be obtained, which may include information such as hourly or daily power generation and power generation power curve. To obtain historical weather data, historical weather data can be obtained from the Meteorological Bureau or other meteorological data providers, including information such as temperature, wind speed, and sunshine hours in the past period of time. To obtain future weather data, weather forecast data for a period of time in the future can also be obtained from the Meteorological Bureau or other meteorological data providers. By analyzing historical output data and historical weather data, a model can be established to predict the future power generation of distributed power sources. Combined with future weather data, future power supply conditions can be predicted more accurately, which helps to prepare for power restoration in advance. By analyzing historical and future weather data, the performance of distributed power sources under different weather conditions can be better understood, which helps to optimize the design and operation of the power supply system and improve its ability to adapt to different weather conditions.

[0034] Step S104, using the pre-trained neural network model, based on the historical output data, historical weather data and future weather data, predict the future output of the distributed power source to obtain a future output prediction result.

[0035] The pre-trained neural network model is used to represent a pre-trained neural network model for predicting the future output of distributed power sources.

[0036] The above-mentioned pre-trained neural network model may refer to a neural network model that has been trained on a data set. The pre-trained neural network model can learn the potential features and patterns in the data by training on the data set, thereby improving the generalization ability and prediction accuracy of the model. The use of pre-trained neural network models can help to better understand the characteristics and change patterns of distributed power sources, thereby improving the accuracy of predictions for future output. The pre-trained neural network model can be a deep neural network, a recurrent neural network, or an attention mechanism network. In specific applications, a suitable pre-trained neural network model can be selected according to the complexity of the problem and the characteristics of the data to meet the needs of the distribution network power supply restoration field.

[0037] In an optional embodiment, a suitable neural network structure, such as a recurrent neural network or a long short-term memory network, can be selected, and the neural network model can be trained based on historical output data and weather data. In the pre-training stage, an unsupervised learning algorithm or a self-supervised learning algorithm can be used for training. Then, the pre-trained neural network model can be fine-tuned and optimized to improve the accuracy of the prediction of future output, and a supervised learning algorithm, such as a back-propagation algorithm, can be used to adjust the model. Finally, the optimized neural network model can be used in combination with future weather data to predict the future output of distributed power sources and obtain future output prediction results. By predicting the future output of distributed power sources and cooperating with power grid dispatch, power supply can be arranged more accurately and the efficiency of power supply restoration can be improved. By predicting future output in advance, it is possible to better cope with uncertain weather factors, reduce fluctuations in the power system, and improve system stability and reliability.

[0038] Step S106: Based on the future output prediction result, the distributed power source is subjected to load shedding operation to obtain a power supply restoration result.

[0039] The power supply restoration result is used to indicate the power supply restoration result obtained after the load shedding operation is performed on the distributed power source.

[0040] The above-mentioned load shedding operation can refer to temporarily reducing part of the load demand in the distribution network to cope with the situation of tight or insufficient power supply. For example, it can be implemented in the following ways. For industrial enterprises, power consumption can be reduced by adjusting production plans or reducing the operating time of certain production lines. At the same time, energy-saving measures can also be taken, such as optimizing equipment operating parameters, improving process flow, etc., to improve energy efficiency and reduce power demand. Through a reasonable load shedding strategy, the reliability and flexibility of power supply can be effectively improved.

[0041] In an optional embodiment, by analyzing the future output prediction results, it can be determined to perform corresponding load shedding operations on distributed power sources. Distributed power sources that are determined to need to perform load shedding operations can be scheduled, that is, load switching or adjustment can be performed at an appropriate time point, and factors such as network load balance, grid stability, and distributed power supply operation efficiency can be considered. Finally, the power supply recovery result after the load shedding operation can be executed, that is, the recovery status and effect of the power supply system can be checked. By predicting future load demand and distributed power supply output in advance, the load of distributed power sources can be flexibly adjusted, so as to better cope with different power supply demands and network conditions, improve the reliability and stability of the power supply system, and effective power supply recovery results can ensure that users can be restored in time when power outages or failures occur, guarantee users' power needs, and improve user experience and satisfaction.

[0042] In an embodiment of the present invention, a power supply restoration method for a distribution network is provided, comprising: obtaining historical output data, historical weather data and future weather data of distributed power sources in the distribution network, wherein the distributed power sources include at least photovoltaic power generation components and wind power generation components; using a pre-trained neural network model, based on the historical output data, historical weather data and future weather data, predicting the future output of the distributed power sources to obtain a future output prediction result, wherein the pre-trained neural network model is used to represent a pre-trained neural network model for predicting the future output of the distributed power sources; based on the future output prediction result, performing a load shedding operation on the distributed power sources to obtain a power supply restoration result, wherein the power supply restoration result is used to represent a power supply restoration result obtained after the load shedding operation is performed on the distributed power sources. It is easy to notice that the present application can predict future output in a relatively short time by acquiring historical output data, historical weather data and future weather data, and using a pre-trained neural network model to predict future output, thereby quickly formulating load shedding operation strategies and improving the efficiency of power supply restoration. By more accurately predicting future output conditions, excessive or insufficient power supply restoration operations can be avoided, thereby reducing the cost of power supply restoration. At the same time, it can better cope with the volatility and uncertainty of distributed power sources in the distribution network, improve the reliability and stability of the power supply system, and thus solve the technical problem of low efficiency of power supply restoration for distribution networks in related technologies.

[0043] Optionally, a pre-trained neural network model is used to predict the future output of distributed power sources based on historical output data, historical weather data and future weather data to obtain future output prediction results, including: determining the degree of overlap between future weather and historical weather of distributed power sources based on historical weather data and future weather data; and obtaining future output prediction results based on the degree of overlap, historical output data, historical weather data and future weather data using a pre-trained neural network model.

[0044] The above-mentioned coincidence degree may refer to a plurality of coincidence degrees determined based on historical weather data and future weather data. Specifically, different coincidence degrees may be determined based on historical weather data and future weather data at a plurality of different moments.

[0045] In an optional embodiment, based on historical weather data and future weather data, the overlap between the future weather of the distributed power source and the historical weather can be determined, which can be achieved by calculating the similarity between various meteorological conditions. For example, the correlation between historical weather data and future weather data can be calculated to determine the overlap between them. Then, using the pre-trained neural network model, based on the overlap, historical output data, historical weather data and future weather data, the prediction result of the future output can be obtained. The future output prediction result can help the distribution network operator make corresponding adjustments and decisions during the power supply restoration process to ensure the stable power supply of the power grid. By using the pre-trained neural network model, the accuracy and reliability of the future output of the distributed power source can be improved based on the overlap, historical output data, historical weather data and future weather data. By considering the changes in weather data and the regularity of historical data, the future output can be better predicted, helping the distribution network to achieve power supply restoration and improving the stability and reliability of power supply.

[0046] Optionally, a pre-trained neural network model is used to obtain future output prediction results based on the overlap, historical output data, historical weather data and future weather data, including: determining the future weather prediction value of the distributed power source based on the overlap and preset conditions, wherein the preset conditions are used to represent pre-set judgment conditions for measuring the overlap; and obtaining future output prediction results based on the overlap, preset conditions, historical output data, historical weather data and future weather prediction values ​​using a pre-trained neural network model.

[0047] The above-mentioned preset conditions can be set to determine whether a certain degree of overlap is greater than the average value of all degrees of overlap. In practical applications, it can be set that when a certain degree of overlap is greater than the average value of all degrees of overlap, that is, historical data in which the degree of overlap is greater than the average value can be selected. The preset conditions can also be determined according to actual needs and are not limited here.

[0048] In an optional embodiment, the collected data may be subjected to feature extraction, including overlap, historical output data, historical weather data, and future weather data, and the overlap may be obtained by calculating the correlation between the historical output data and the historical weather data. Next, a predetermined judgment condition for measuring the overlap may be set, such as setting a correlation threshold, etc. The trained neural network model may be used to input the overlap, preset conditions, historical output data, historical weather data, and future weather forecast values ​​to obtain a forecast result of future output. Through the above steps, by setting preset conditions and using the overlap for judgment, the credibility of the forecast result may be better evaluated, and the robustness and stability of the system may be improved.

[0049] Optionally, a pre-trained neural network model is used to obtain future output prediction results based on the degree of overlap, preset conditions, historical output data, historical weather data and future weather forecast values, including: based on the degree of overlap and preset conditions, historical output data and historical weather data are screened to obtain target historical output data and target historical weather data; and the target historical output data, target historical weather data and future weather forecast values ​​are used as input variables of the pre-trained neural network model to obtain future output prediction results.

[0050] In an optional embodiment, the target historical output data and the target historical weather data can be obtained by screening according to the overlap and preset conditions. Then, the target historical output data, the target historical weather data and the future weather forecast value can be used as input variables and input into the pre-trained neural network model. In the neural network model, by learning the patterns of historical output data and historical weather data, as well as the impact of future weather forecast values ​​on the output, the forecast results of future output can be obtained. By screening historical data based on the overlap and preset conditions, data with little impact on the forecast results can be removed, thereby improving the efficiency and accuracy of the model. Combining historical data with future forecast data can more comprehensively consider historical trends and future changes, making the forecast results more accurate and reliable.

[0051] Optionally, the method further includes: obtaining an initial neural network model; and training the initial neural network model based on future weather forecast values ​​to obtain a pre-trained neural network model.

[0052] The above-mentioned initial neural network model may refer to a general or untrained neural network model. The initial neural network model may be selected according to actual needs and is not limited here.

[0053] In an optional embodiment, the structure and parameters of the neural network can be determined, and the neural network structure can be selected, such as a multi-layer perceptron and a convolutional neural network. Then, the parameters such as the input, output and number of nodes in the hidden layer of the network can be determined according to the specific application scenario and requirements. Then, the initial neural network model can be trained based on the future weather forecast value, and the historical weather data and the power supply recovery data of the distribution network are used as the training set, the future weather forecast value is used as the input, and the power supply recovery data of the distribution network is used as the output. The weight and bias of the neural network are continuously adjusted through the back propagation algorithm so that the prediction result of the network is as close as possible to the actual result. After multiple rounds of training, the weight and bias parameters of the neural network will gradually converge to obtain a pre-trained neural network model. This model can be used in the tasks of future weather forecasting and power supply recovery of the distribution network to improve the accuracy and reliability of the system. By using the future weather forecast value for training, the generalization ability of the neural network model can be improved, so that the model can make accurate power supply recovery predictions under unknown weather conditions.

[0054] Optionally, based on historical weather data and future weather data, the degree of overlap between future weather and historical weather of distributed power sources is determined, including: determining the amount of change in historical weather data based on historical data at different times in the historical weather data; constructing the amount of change in future weather data based on a target number of samples and future weather data, wherein the target number of samples is used to represent the number of samples of future weather data; determining the degree of overlap based on the amount of change in historical weather data and the amount of change in future weather data.

[0055] In an optional embodiment, the amount of change in historical weather data, such as the amount of change in temperature, the amount of change in humidity, etc., can be calculated through the historical data at different times in the historical weather data. Then, the amount of change in future weather data can be constructed based on the target number of samples and the future weather data to predict the changing trend of future weather. Finally, based on the amount of change in historical weather data and the amount of change in future weather data, the degree of overlap between future weather and historical weather can be determined, thereby judging the reliability and accuracy of future weather data. By determining the degree of overlap, the power supply strategy can be adjusted in a timely manner to reduce power supply interruptions and failures caused by weather changes, thereby improving power supply reliability and efficiency.

[0056] The technical solution proposed in this application is explained below in combination with an optional application scenario. The purpose of this application is to provide a method and system for power supply restoration after an accident in a distribution network containing distributed power sources. First, the historical output data, historical weather data and future weather data of wind and photovoltaic power generation components are obtained, and the overlap between future weather and historical weather is calculated. Then, based on the overlap and preset conditions, the future weather forecast value is obtained, and the future weather forecast value is used to train the traditional neural network model. The future output forecast result of the wind and photovoltaic power generation components is obtained according to the trained neural network model. Finally, based on the future output forecast result, the load shedding operation is performed and the wind and photovoltaic power generation components are used to restore power. The present application may only require a small amount of historical output data, historical weather data and future weather data, avoiding the construction of a complex power restoration model and the execution of cumbersome algorithms. At the same time, the wind and photovoltaic output can be predicted more accurately based on the change in weather characteristics.

[0057] The present application proposes a method for restoring power supply in the event of an accident in a distribution network containing distributed power sources, wherein the distributed power sources include photovoltaic power generation components and wind power generation components, and the restoration method includes the following steps: obtaining historical output data, historical weather data, and future weather data of the photovoltaic power generation components and the wind power generation components; predicting the future output of the photovoltaic power generation components and the wind power generation components based on the historical output data, historical weather data, and future weather data; performing load shedding operations based on the future output prediction results; and utilizing the photovoltaic power generation components and the wind power generation components to supply power to the distribution network load.

[0058] Optionally, future weather data information includes wind speed, wind direction, light intensity, light duration, and temperature; historical weather data includes wind speed, wind direction, light intensity, light duration, and temperature. Future weather data information includes wind speed, wind direction, light intensity, light duration, and temperature. Future weather data can be obtained through weather observation stations or weather satellites, radars and other weather forecast data sources; historical weather data includes wind speed, wind direction, light intensity, light duration, and temperature. Historical weather data is usually stored in a corresponding database with data entries constructed with fields such as time point, weather data, storage time, sensing device number, and sensing device geographic location, and can be obtained by accessing the historical weather database.

[0059] Optionally, historical wind speed changes, historical wind direction changes, historical light intensity changes, historical light duration changes, and historical temperature changes can be constructed based on the wind speed, wind direction, light intensity, light duration, and temperature in the historical data; the number of future weather data samples is set, and the future wind speed changes, future wind direction changes, future light intensity changes, future light duration changes, and future temperature changes are constructed; the overlap between future weather and historical weather is calculated; based on preset conditions and the overlap, future weather forecast values ​​are obtained; the obtained future weather forecast values ​​are used to train the neural network model, and the historical output data and historical weather data in the historical data whose overlap meets the preset conditions, as well as the future weather forecast values, are used as input variables, and are input into the trained neural network model to obtain future output forecast results for photovoltaic power generation components and wind power generation components.

[0060] Optionally, the wind speed, wind direction, light intensity, light duration, and temperature at the x+2th time point and the xth time point in the historical data can be taken, and the historical data at the x+2th time point can be subtracted from the historical data at the xth time point to obtain the corresponding historical wind speed change, historical wind direction change, historical light intensity change, historical light duration change, and historical temperature change; for example, when the time interval is set to 1 hour, the historical data at 13:00 and the historical data at 15:00 can be taken, and the wind speed in the historical data at 15:00 can be subtracted from the wind speed in the historical data at 13:00 to obtain a historical wind speed change, and the wind speed in the historical data at 16:00 can be subtracted from the wind speed in the historical data at 14:00 to obtain the next historical wind speed change, and similarly, the historical wind direction change, historical light intensity change, historical light duration change, and historical temperature change can be obtained. Of course, you can also collect data and calculate the changes in related historical data at time intervals of 15 minutes, 30 minutes, 12 hours, 1 day, 2 days, etc. There is no restriction here.

[0061] Set the number of future weather data samples, and construct the future wind speed change, future wind direction change, future light intensity change, future light duration change, and future temperature change; set the wind speed, wind direction, light intensity, light duration, and temperature at the y+2th time point and the yth time point in the future data, and subtract the future data at the yth time point from the y+2th time point to obtain the corresponding future wind speed change, future wind direction change, future light intensity change, future light duration change, and future temperature change. The future weather data change can be obtained by a method similar to the above method of obtaining the historical data change.

[0062] Optionally, the degree of overlap between future weather and historical weather can be calculated according to the following formula:

[0063]

[0064] Among them, C m represents the overlap between future weather and historical weather, k represents the number of changes, n represents the total number of changes, WS k represents the kth future wind speed change, WSH k Indicates the kth historical wind speed change, WD k Indicates the kth future wind direction change, WDH k represents the kth historical wind direction change, LI k Indicates the kth future light intensity change, LIH k Indicates the kth historical light intensity change, LT k Indicates the change in the kth future illumination duration, LTH k Indicates the change in the kth historical illumination duration, T k represents the kth future temperature change, TH k Represents the kth historical temperature change.

[0065] Optionally, the future weather forecast value can be obtained based on the preset conditions and the coincidence degree. First, the calculated coincidence degree C is determined. m The maximum value C mmax and minimum value C mmin , and find the maximum value C mmax and minimum value C mmin The average value of the two mave , the preset condition is the overlap degree C m Greater than average C mave , that is, select the overlap degree C m Greater than average C mave For example, when the maximum value C mmax and minimum value C mmin When the average C mave = (0.88+0.2) / 2=0.54, which satisfies the above preset condition, namely, the overlap degree C m Greater than 0.54. The calculation formula for future weather forecast values ​​is as follows:

[0066]

[0067] Among them, WS d+1 represents the wind speed forecast value at the future time d+1, d = 0, 1, 2, 3..., WS0 represents the wind speed at the current time, WSH sd Indicates the historical wind speed change at time d in the historical data where the coincidence degree is greater than the average value, C s Indicates the sth coincidence degree in the historical data where the coincidence degree is greater than the average value, WD d+1It indicates the predicted wind direction at the future time d+1, WD0 indicates the wind direction at the current time, and WDH sd Indicates the historical wind direction change at time d in the historical data where the coincidence degree is greater than the average value, LI d+1 represents the predicted light intensity at the future time d+1, LI0 represents the light intensity at the current time, LIH sd Indicates the historical light intensity change at time d in the historical data where the coincidence degree is greater than the average value, LT d+1 It indicates the predicted value of the illumination duration at the future time d+1, LT0 indicates the illumination duration at the current time, LTH sd Indicates the change in historical illumination duration at time d in the historical data where the coincidence degree is greater than the average value, T d+1 represents the predicted temperature at time d+1 in the future, T0 represents the temperature at the current time, TH sd Indicates the sth historical temperature change at time d in the historical data with a coincidence greater than the average value, l indicates that the coincidence degree is greater than the average value C mave The total number of historical data, for example, when the average value C mave When the overlap degree is 0.54, select C m If the historical data set is greater than 0.54, such as the historical data set includes historical data of 15 moments, then l = 15. It should be noted that the above moments can be at intervals of 15 minutes, 30 minutes, 1 hour, 12 hours, etc., but the historical weather data and future weather data need to match the above intervals, that is, if the future weather data is calculated at intervals of 1 hour, the historical weather data obtained should also be at intervals of 1 hour.

[0068] Optionally, the obtained future weather forecast value can be used to train the neural network model, and the historical output data and historical weather data in the historical data whose overlap meets the preset conditions, as well as the future weather forecast value, are used as input variables and input into the trained neural network model to obtain the future output forecast results of the photovoltaic power generation components and the wind power generation components. Here, the neural network model is first selected as one of the convolutional neural network model, the LSTM neural network model or the recurrent neural network model, and then the future weather forecast value is input into the neural network model for training. The number of training times is preferably about 20 times. Finally, the historical output data, historical weather data and future weather forecast value whose overlap meets the preset conditions are input as input variables into the above-mentioned trained neural network model to obtain the future output forecast results of wind and solar power.

[0069] Optionally, the load shedding operation in the present application may include: recording the output prediction value of the photovoltaic power generation component at the d+1th moment as PV predic-d+1 , the predicted output value of the wind power generation component at the d+1th moment is WE predic-d+1: When PV predic-d+1 > PV ref and WE predic-d+1 > WE ref , no load shedding operation is performed, and the photovoltaic power generation components and wind power generation components are directly used to supply power to the distribution network load, where PV ref + WE ref = K1P load ; when PV predic-d+1 ≤ PV ref and WE predic-d+1 > WE ref , or PV predic-d+1 > PV ref and WE predic-d+1 ≤ WE ref , the tertiary interruptible load is removed and the photovoltaic power generation components and wind power generation components are used to supply power to the remaining load of the distribution network; when PV predic-d+1 < PV ref and WE predic-d+1 < WE ref , the tertiary interruptible load and the secondary interruptible load are removed and the photovoltaic power generation components and wind power generation components are used to supply power to the remaining load of the distribution network; where PV ref represents the reference value of photovoltaic output, WE ref represents the reference value of wind power output, P load represents the maximum load demand of the distribution network, and K1 represents the weighting coefficient, 1 < K1 < 2. The interruptible load may refer to the load that can be interrupted under the power grid outage or emergency state by signing an economic contract or economic agreement with the power company. Among them, the economic loss after removing the tertiary interruptible load is less than the economic loss after removing the secondary interruptible load, and the economic loss after removing the secondary interruptible load is less than the economic loss after removing the primary interruptible load.

[0070] Optionally, the neural network model can be any one of a convolutional neural network model, a neural network model (LSTM), or a recurrent neural network model.

[0071] This application also proposes a distribution network accident power supply restoration system with distributed power sources, which is used to execute the above-mentioned distribution network accident power supply restoration method with distributed power sources. The system includes a data acquisition module, an output prediction module, and a load operation module. The acquisition module is used to acquire the historical output data, historical weather data, and future weather data of the photovoltaic power generation components and wind power generation components. The output prediction module is used to predict the future output of the photovoltaic power generation components and wind power generation components based on the historical output data, historical weather data, and future weather data. The load operation module is used to perform load shedding operations based on the future output prediction results, and finally use the photovoltaic power generation components and wind power generation components to supply power to the distribution network load.

[0072] The beneficial effects of the present application are as follows: only a small amount of historical output data, historical weather data and future weather data need to be obtained to realize wind and solar power output prediction, avoiding the construction of complex power models and the execution of cumbersome algorithms; introducing the change in weather characteristics into the wind and solar power output prediction, better fitting the characteristics of wind and solar power output fluctuating with weather characteristics, thus enabling more accurate wind and solar power output prediction.

[0073] Figure 2 is a flow chart of an optional power supply restoration method of a distribution network according to an embodiment of the present invention, such as Figure 2 As shown, it includes: S202, obtaining the historical output data, historical weather data and future weather data of the distributed power sources in the distribution network. S204, based on the historical weather data and the future weather data, determining the overlap between the future weather and the historical weather of the distributed power sources. S206, based on the overlap and the preset conditions, determining the future weather forecast value of the distributed power sources. S208, using the pre-trained neural network model, based on the overlap, the preset conditions, the historical output data, the historical weather data and the future weather forecast value, obtaining the future output forecast result. S210, based on the future output forecast result, performing load shedding operation on the distributed power sources to obtain the power supply recovery result.

[0074] According to another aspect of an embodiment of the present invention, a power supply restoration device for a distribution network is also provided. The device can execute the power supply restoration method for the distribution network of the above embodiment. The specific implementation method and preferred application scenario are the same as the above embodiment and will not be repeated here.

[0075] Figure 3 is a schematic diagram of a power supply restoration device for a distribution network according to an embodiment of the present application, such as Figure 3 As shown, the device includes the following: an acquisition module 302 , a prediction module 304 and an execution module 306 .

[0076] Among them, the acquisition module is used to obtain the historical output data, historical weather data and future weather data of the distributed power sources in the distribution network, wherein the distributed power sources include at least photovoltaic power generation components and wind power generation components; the prediction module is used to use the pre-trained neural network model to predict the future output of the distributed power sources based on the historical output data, historical weather data and future weather data, and obtain the future output prediction result, wherein the pre-trained neural network model is used to represent the pre-trained neural network model for predicting the future output of the distributed power sources; the execution module is used to perform a load shedding operation on the distributed power sources based on the future output prediction result, and obtain a power supply recovery result, wherein the power supply recovery result is used to represent the power supply recovery result obtained after the load shedding operation is performed on the distributed power sources.

[0077] Among them, the prediction module is also used to determine the overlap between the future weather and historical weather of the distributed power source based on historical weather data and future weather data; and to obtain the future output prediction result based on the overlap, historical output data, historical weather data and future weather data by using the pre-trained neural network model.

[0078] Among them, the prediction module is also used to determine the future weather forecast value of the distributed power source based on the overlap and preset conditions, wherein the preset conditions are used to represent the pre-set judgment conditions for measuring the overlap; and the future output forecast result is obtained based on the overlap, preset conditions, historical output data, historical weather data and future weather forecast values ​​by using the pre-trained neural network model.

[0079] Among them, the prediction module is also used to screen the historical output data and historical weather data based on the overlap and preset conditions to obtain the target historical output data and target historical weather data; the target historical output data, target historical weather data and future weather forecast values ​​are used as input variables of the pre-trained neural network model to obtain the future output forecast results.

[0080] Among them, the prediction module is also used to obtain an initial neural network model; based on future weather forecast values, the initial neural network model is trained to obtain a pre-trained neural network model.

[0081] Among them, the prediction module is also used to determine the change in historical weather data based on historical data at different times in the historical weather data; construct the change in future weather data based on the target number of samples and future weather data, wherein the target number of samples is used to represent the number of samples of future weather data; and determine the degree of overlap based on the change in historical weather data and the change in future weather data.

[0082] An embodiment of the present application further provides an electronic device, comprising: a memory storing an executable program; and a processor for running the program, wherein the program executes the methods in various embodiments of the present invention when running.

[0083] The above-mentioned memory may refer to a device inside a computer for storing data and programs, and may include memory, hard disk, etc., wherein the memory may be used to temporarily store running programs and data, the hard disk may be used to store programs and data for a long time, and the memory may be used to enable the computer to read and write data, and execute programs; the above-mentioned processor may be responsible for executing instructions in computer programs and performing data processing, and may be responsible for controlling and executing various operations, including arithmetic operations, logical operations, data transmission, etc.

[0084] An embodiment of the present application further provides a computer-readable storage medium, which includes a stored executable program, wherein when the executable program is running, the device where the computer-readable storage medium is located is controlled to execute the methods in various embodiments of the present invention.

[0085] The above-mentioned computer storage medium may refer to a medium in a computer memory used to store certain discontinuous physical quantities. Computer storage media mainly include semiconductors, magnetic cores, magnetic drums, magnetic tapes, laser disks, etc.; the stored program included in the computer-readable storage medium may be a set of instructions that can be recognized and executed by a computer, running on an electronic computer, and is an information tool that meets certain needs of people.

[0086] An embodiment of the present application further provides a computer program product, including a computer program, which implements the methods in various embodiments of the present invention when executed by a processor.

[0087] The above-mentioned computer program product may refer to a software program that has been written, tested and released and can be run on a computer or other device. The computer program product may include an application, an operating system, tool software, etc., which is used to implement specific functions or solve specific problems.

[0088] An embodiment of the present application further provides a computer program product, including a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium is used to store a computer program, and when the computer program is executed by a processor, the method in each embodiment of the present invention is implemented.

[0089] The above-mentioned non-volatile computer-readable storage medium may refer to a medium for storing data. The non-volatile computer-readable storage medium can keep the data from being lost when the power is off, and can be used to store long-term data, such as operating systems, applications, and user files. The non-volatile storage medium may include hard disk drives, solid-state drives, optical disks, and flash memory storage devices, etc.

[0090] The embodiments of the present application further provide a computer program, which implements the methods in the above-mentioned embodiments of the present invention when executed by a processor.

[0091] The above-mentioned computer program may refer to a collection of instructions used to tell a computer to perform a specific task or operation. A computer program may be written by a programmer using a specific programming language and may include algorithms, data structures, logic, and control flows. A computer program may be used for a variety of purposes, including application software, operating systems, and the like.

[0092] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0093] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0094] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0095] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0096] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0097] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A power supply restoration method for a distribution network, characterized in that: include: Acquire historical output data, historical weather data and future weather data of distributed power sources in the distribution network, wherein the distributed power sources include at least photovoltaic power generation components and wind power generation components; Using a pre-trained neural network model, based on the historical output data, historical weather data and future weather data, the future output of the distributed power source is predicted to obtain a future output prediction result, wherein the pre-trained neural network model is used to represent a pre-trained neural network model for predicting the future output of the distributed power source; Based on the future output prediction result, a load shedding operation is performed on the distributed power source to obtain a power supply recovery result, wherein the power supply recovery result is used to represent a power supply recovery result obtained after the load shedding operation is performed on the distributed power source.

2. The power supply restoration method of the distribution network according to claim 1, characterized in that: Using the pre-trained neural network model, based on the historical output data, historical weather data and future weather data, the future output of the distributed power source is predicted to obtain a future output prediction result, including: Based on the historical weather data and the future weather data, determining the degree of overlap between the future weather of the distributed power source and the historical weather; The pre-trained neural network model is used to obtain the future output prediction result based on the overlap, the historical output data, the historical weather data and the future weather data.

3. The power supply restoration method of the distribution network according to claim 2, characterized in that: The pre-trained neural network model is used to obtain the future output prediction result based on the overlap, the historical output data, the historical weather data and the future weather data, including: Determining a future weather forecast value of the distributed power source based on the overlap and a preset condition, wherein the preset condition is used to represent a pre-set determination condition for measuring the overlap; The pre-trained neural network model is used to obtain the future output prediction result based on the overlap, the preset conditions, the historical output data, the historical weather data and the future weather forecast value.

4. The power supply restoration method of the distribution network according to claim 3, characterized in that: The pre-trained neural network model is used to obtain the future output prediction result based on the overlap, the preset conditions, the historical output data, the historical weather data and the future weather forecast value, including: Based on the overlap degree and the preset condition, the historical output data and the historical weather data are screened to obtain target historical output data and target historical weather data; The target historical output data, the target historical weather data and the future weather forecast value are used as input variables of the pre-trained neural network model to obtain the future output forecast result.

5. The power supply restoration method of the distribution network according to claim 3, characterized in that: The method further comprises: Get the initial neural network model; Based on the future weather forecast value, the initial neural network model is trained to obtain the pre-trained neural network model.

6. The power supply restoration method of the distribution network according to claim 2, characterized in that: Determining the degree of overlap between future weather and historical weather of the distributed power source based on the historical weather data and the future weather data includes: Determine the amount of change in the historical weather data based on the historical data at different times in the historical weather data; Based on the target number of samples and the future weather data, constructing a change amount of future weather data, wherein the target number of samples is used to represent the number of samples of the future weather data; The degree of overlap is determined based on the amount of change in the historical weather data and the amount of change in the future weather data.

7. A power supply restoration device for a distribution network, characterized in that: include: An acquisition module, used to acquire historical output data, historical weather data and future weather data of distributed power sources in the distribution network, wherein the distributed power sources include at least photovoltaic power generation components and wind power generation components; A prediction module, used to use a pre-trained neural network model to predict the future output of the distributed power source based on the historical output data, historical weather data and future weather data, and obtain a future output prediction result, wherein the pre-trained neural network model is used to represent a pre-trained neural network model for predicting the future output of the distributed power source; An execution module is used to perform a load shedding operation on the distributed power source based on the future output prediction result to obtain a power supply recovery result, wherein the power supply recovery result is used to represent the power supply recovery result obtained after performing the load shedding operation on the distributed power source.

8. An electronic device, characterized in that: include: A memory storing an executable program; A processor is used to run the program, wherein when the program is run, the power supply restoration method for the distribution network described in any one of claims 1 to 6 is executed.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored executable program, wherein when the executable program is executed, the device where the storage medium is located is controlled to execute the power supply restoration method for the distribution network according to any one of claims 1 to 6.

10. A computer program product, characterized in that The invention comprises a computer program, which, when executed by a processor, implements the power supply restoration method for a distribution network according to any one of claims 1 to 6.

Images