New energy station management system and method based on smart power grid

Abstract

The invention discloses a new energy station management system and method based on an intelligent power grid, and the system comprises a real-time monitoring module which is used for predicting the charging load fluctuation of an electric vehicle based on real-time monitoring data, a scheduling module which is used for adjusting the output of new energy power generation equipment according to the predicted load fluctuation, and an optimization module. The power distribution network management module is used for optimizing charging scheduling in a charging station and avoiding overload caused by peak charging at the same time, the dynamic adjustment module is used for dynamically adjusting the voltage and frequency of a power distribution network, and the real-time monitoring module is used for monitoring the state of the power distribution network in real time and dynamically updating a management strategy; according to the new energy station management system and method based on the smart power grid, the problems that in the prior art, due to instability and rapid change characteristics of electric vehicle charging loads, severe fluctuation of the loads can bring extra pressure to a power distribution network, the overload problem of the power distribution network is caused, and then the reliability and safety of the power grid are affected are solved.

Description

Technical Field

[0001] The present invention relates to the field of smart grid technology, and in particular to a new energy station management system and method based on a smart grid. Background Art

[0002] In modern energy systems, with the large-scale access of renewable energy (such as wind and solar energy), the management and dispatch of power grids are facing increasingly complex challenges. The new energy site management method based on smart grids has achieved the optimal management of renewable energy by introducing advanced technologies such as big data, cloud computing and the Internet of Things. The application of these technologies can not only improve the real-time monitoring and dispatching capabilities of renewable energy such as wind and solar energy, but also effectively improve the overall operating efficiency and stability of the power system.

[0003] The patent with publication number CN118769969A discloses a high-efficiency charging device for a stereoscopic parking garage and a control method thereof. By collecting and preprocessing vehicle distribution, charging status and grid load data, high-precision prediction of vehicle distribution is achieved. Therefore, although the existing new energy station management has certain advantages in renewable energy integration and distribution, there are still obvious problems in dealing with large-scale electric vehicle charging loads. Due to the instability and rapid change characteristics of electric vehicle charging loads, the drastic fluctuations of such loads will bring additional pressure to the distribution network, especially during the peak period of charging demand, resulting in a significant increase in the load fluctuations of the distribution network. Such load fluctuations will not only increase the difficulty of grid scheduling, but also cause overload problems in the distribution network, thereby affecting the reliability and safety of the grid. Summary of the invention

[0004] The purpose of the present invention is to provide a new energy station management system and method based on smart grid, so as to solve the problem in the prior art that due to the instability and rapid change characteristics of electric vehicle charging load, the drastic fluctuation of such load will bring additional pressure to the distribution network, especially during the peak period of charging demand, resulting in significantly increased load fluctuations in the distribution network. Such load fluctuations will not only increase the difficulty of grid scheduling, but also cause overload problems in the distribution network, thereby affecting the reliability and safety of the grid.

[0005] To achieve the above object, the present invention provides the following technical solution: a new energy station management system based on smart grid, the system comprising:

[0006] A prediction module, used to predict the fluctuation of electric vehicle charging load based on real-time monitoring data;

[0007] The dispatching module is used to adjust the output of renewable energy power generation equipment according to the predicted load fluctuations;

[0008] The optimization module is used to optimize the charging scheduling in the charging station to avoid overload caused by peak charging at the same time; the optimization module defines a fitness function f(p i ), measure the pros and cons of each scheduling scheme, and calculate the fitness function f(p i ) Select a scheduling plan and exchange scheduling plan information, generate a new scheduling plan, and iterate to find the optimal charging scheduling strategy to avoid peak overload;

[0009] The specific process of exchanging the information of the scheduling scheme and generating a new scheduling scheme is as follows:

[0010] p new =αp i +(1-α)p j ,

[0011] Among them, p new represents the new scheduling scheme, α represents the crossover ratio, and p i and p j Indicates two different scheduling schemes;

[0012] A dynamic adjustment module, used to dynamically adjust the voltage and frequency of the power distribution network;

[0013] The dynamic adjustment module records the real-time values ​​of the voltage and frequency of the power grid V t and F t , according to the ideal voltage mean μ of the power grid operation V and the ideal frequency mean μ F , real-time monitoring of the voltage and frequency deviation ΔV=V t -μ V and ΔF=F t -μ F , set a threshold value, if the deviation exceeds the threshold value, adjust the power generation equipment and load distribution in the power grid, and gradually increase the voltage V t and frequency F t Adjust back to mean μ V and μ F Department;

[0014] The real-time monitoring module is used to monitor the status of the power distribution network in real time and dynamically update the management strategy.

[0015] Preferably, the real-time monitoring module predicts the fluctuation of charging load of electric vehicles based on the real-time monitoring data, including:

[0016] Based on the historical charging load data, the support vector machine is used to find an optimal hyperplane to divide these data into different categories. The support vector machine finds the optimal hyperplane that separates these data, specifically:

[0017] When satisfied Theni (w·x i +b)≥1,i=1,2,…,

[0018] That is, if for every data point x i , its classification label y i Satisfy i (w·x i +b)≥1, indicating that the data point is correctly classified and is outside the correct category boundary;

[0019] If any data point x i Not satisfied with y i (w·x i +b)≥1, the point is misclassified or falls on the boundary of the hyperplane, and the normal vector w and offset b of the hyperplane will continue to be optimized so that all points meet the conditions;

[0020] Among them, w represents the normal vector of the hyperplane, b represents the offset, i represents the index variable, and ||w|| represents the Euclidean norm of the vector w;

[0021] The final classification function is f(x), which is used to determine the category of load fluctuation. If the result of function f(x) is +1, it means that the current charging load is within the normal range. If f(x) is -1, it means that an abnormal load fluctuation is detected.

[0022] Preferably, the dynamic adjustment module dynamically adjusts the voltage and frequency of the power distribution network further includes:

[0023] The mean regression formula for voltage is:

[0024] V t+1 =V t +κ(μ V -V t ),

[0025] Among them, V t+1 represents the voltage value at the next moment, κ represents the regression rate constant, V t Indicates the voltage value at the current moment, μ V Represents the ideal voltage mean.

[0026] Preferably, the real-time monitoring module includes:

[0027] The high-dimensional features of the power grid state are extracted based on the deep belief network, and the energy function is used to characterize the correlation between input and output, specifically:

[0028]

[0029] Where E(v, h) represents the energy function, v i represents the input unit of the visible layer, hj represents the hidden layer unit, b i represents the bias parameter of the visible layer, c j represents the bias parameter of the hidden layer, and i and j both represent index variables.

[0030] Preferably, the scheduling module can transfer the state of the power grid to a new state according to the current action, and obtain a reward after taking a certain action.

[0031] Preferably, the optimization module exchanges the information of the scheduling scheme and generates a new scheduling scheme in the following specific process:

[0032] p new =αp i +(1-α)p j ,

[0033] Among them, p new represents the new scheduling scheme, α represents the crossover ratio, and p i and p j Indicates two different scheduling schemes.

[0034] Preferably, the real-time monitoring module also includes collecting historical charging data and current charging demand of each charging station, analyzing these data, extracting charging behavior patterns, predicting future charging load fluctuations, and generating charging load fluctuation warnings based on the prediction results.

[0035] Preferably, the collection of historical charging data and current charging demand of each charging station specifically includes obtaining the historical charging data of each charging station, including the daily charging time and charging amount, collecting the current charging demand of each charging station in real time, including the number of ongoing charging times and the number of reserved charging times, and calculating the charging demand intensity of each charging station based on the historical data and current demand. When the calculated demand intensity I exceeds the preset threshold value T, a demand peak warning is issued, and the formula is: I>T, where I represents the charging demand intensity and T represents the preset threshold value.

[0036] Preferably, the calculation of the charging demand intensity of each charging station based on historical data and current demand includes counting the average daily charging amount of each charging station in the past week, counting the total charging amount currently being charged at each charging station, and calculating the current demand intensity based on the two, the formula is: I=C+(P / D), where I represents the charging demand intensity, C represents the average daily charging amount, P represents the current charging amount, and D represents the remaining time of the day, and the calculation results are stored in a database.

[0037] A new energy station management method based on smart grid, using the new energy station management system based on smart grid, the method comprising:

[0038] S1. Based on real-time monitoring data, predict the charging load fluctuation of electric vehicles. Based on the historical charging load data, find an optimal hyperplane through support vector machine, divide these data into different categories, and use classification function to judge the category of load fluctuation. If the function result is +1, it means that the current charging load is in the normal range. If it is -1, it means that the load fluctuation is abnormal.

[0039] S2. Adjust the output of new energy power generation equipment according to the predicted load fluctuations. According to the current action, the state of the power grid is transferred to a new state, and a reward is obtained after taking the action.

[0040] S3. Optimize the charging scheduling in the charging station to avoid overload caused by simultaneous peak charging. Define a fitness function to measure the pros and cons of each charging scheduling plan. Select a better charging scheduling plan based on the fitness function, and generate a new scheduling plan through crossover operations. Through multiple iterations, find the optimal charging scheduling strategy to avoid peak overload.

[0041] S4. Dynamically adjust the voltage and frequency of the distribution network to ensure stable operation, record the real-time values ​​of the voltage and frequency of the power grid, monitor the deviation of voltage and frequency in real time according to the ideal voltage and frequency averages of the power grid operation, set a threshold, and when the deviation exceeds the threshold, adjust the output and load distribution of the power generation equipment to gradually adjust the voltage and frequency back to the average value;

[0042] S5. Monitor the status of the distribution network in real time and extract high-dimensional features of the power grid status based on the deep belief network. Specifically, the features of the visible layer input units and hidden layer units are extracted through the energy function. The extracted features are used to monitor the power grid status in real time and dynamically update the management strategy to prevent overload.

[0043] It can be seen from the above technical solution that the present invention has the following beneficial effects:

[0044] The new energy station management system and method based on the smart grid can effectively predict the fluctuation of large-scale electric vehicle charging load through real-time monitoring and support vector machine technology, reduce the instability caused by drastic changes in charging demand, and significantly improve the accuracy and response speed of electric vehicle load management. The scheduling module adjusts the output of new energy power generation equipment according to the predicted load fluctuation, the optimization module optimizes the charging scheduling in the charging station to avoid overload caused by simultaneous peak charging, the dynamic adjustment module dynamically adjusts the voltage and frequency of the distribution network, the real-time monitoring module monitors the status of the distribution network in real time, and dynamically updates the management strategy. It can dynamically adjust the output of new energy power generation equipment according to the changes in the electric vehicle charging load, realize more efficient distribution and utilization of renewable energy, and reduce the impact of new Grid fluctuations caused by energy instability can reasonably allocate charging tasks and avoid load overload problems caused by concentrated charging during peak hours, thereby effectively reducing the load pressure on the distribution network, and can quickly adjust the output and load distribution of power generation equipment to avoid the impact of fluctuations on the overall operation of the power grid. It can monitor the complex state of the power grid in real time, dynamically identify potential problems and update management strategies, solving the problem of existing technologies due to the instability and rapid change of electric vehicle charging loads. Such drastic fluctuations in loads will bring additional pressure to the distribution network, especially during the peak period of charging demand, resulting in significantly increased load fluctuations in the distribution network. Such load fluctuations will not only increase the difficulty of grid scheduling, but also cause overload problems in the distribution network, thereby affecting the reliability and safety of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 This is a schematic diagram of module connection of the present invention;

[0046] Figure 2 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0047] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions 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 are within the scope of protection of the present invention.

[0048] like Figure 1 As shown, a new energy station management system based on smart grid, the system includes:

[0049] A prediction module, used to predict the fluctuation of electric vehicle charging load based on real-time monitoring data;

[0050] The dispatching module is used to adjust the output of renewable energy power generation equipment according to the predicted load fluctuations;

[0051] The scheduling module can transfer the state of the power grid to a new state according to the current action, and obtain a reward after taking a certain action;

[0052] The optimization module is used to optimize the charging scheduling in the charging station to avoid overload caused by peak charging at the same time; the optimization module defines a fitness function f(p i ), measure the pros and cons of each scheduling scheme, and calculate the fitness function f(p i ) Select a scheduling plan and exchange scheduling plan information, generate a new scheduling plan, and iterate to find the optimal charging scheduling strategy to avoid peak overload;

[0053] The specific process of exchanging the information of the scheduling scheme and generating a new scheduling scheme is as follows:

[0054] p new =αp i +(1-α)p j ,

[0055] Among them, p new represents the new scheduling scheme, α represents the crossover ratio, and p i and p j Indicates two different scheduling schemes;

[0056] A dynamic adjustment module, used to dynamically adjust the voltage and frequency of the power distribution network;

[0057] The dynamic adjustment module records the real-time values ​​of the voltage and frequency of the power grid V t and F t , according to the ideal voltage mean μ of the power grid operation V and the ideal frequency mean μ F , real-time monitoring of the voltage and frequency deviation ΔV=V t -μ V and ΔF=F t -μ F , set a threshold value, if the deviation exceeds the threshold value, adjust the power generation equipment and load distribution in the power grid, and gradually increase the voltage V t and frequency F t Adjust back to mean μ V and μ F Department;

[0058] The real-time monitoring module is used to monitor the status of the power distribution network in real time and dynamically update the management strategy.

[0059] The core purpose of the new energy station management system is to ensure the stable operation of the new energy station by monitoring, dispatching and optimizing various subsystems, especially in the case of fluctuations in the charging load of electric vehicles. The real-time monitoring module is responsible for collecting the charging load data of electric vehicles in the charging station and making real-time predictions of the load to help other modules effectively respond to load changes. The dispatching module dynamically adjusts the power output of new energy power generation equipment according to the predicted load fluctuations to ensure the balance of power supply and demand. A model based on the state transfer of the power grid is designed internally. By executing dispatch actions and obtaining rewards based on the results, the dispatching strategy is continuously optimized. The optimization module further ensures the efficiency of the dispatching process, especially to avoid overload during peak charging. The module evaluates the effect of each scheduling scheme by setting the fitness function, and generates the optimal scheduling scheme through multiple iterations to reduce the risk of overload. The dynamic adjustment module mainly targets the voltage and frequency changes in the operation of the power grid, monitors and records the voltage and frequency values ​​of the power grid in real time, and dynamically adjusts the output and load distribution of the power generation equipment according to the set ideal mean to ensure the stability of the power grid. The real-time monitoring module monitors the status of the entire distribution network in real time to ensure that the system can respond quickly to abnormal situations and adjust the management strategy to maintain stable operation. By real-time monitoring of the charging load of electric vehicles and dynamically adjusting the output of power generation equipment according to the prediction results, it can more effectively match the supply and demand of electricity and avoid energy waste. The optimization module selects the optimal scheduling scheme through iteration, reduces the overload phenomenon during peak charging, improves the load management capability of the charging station, and enhances the reliability and safety of the system. The dynamic adjustment module ensures that the voltage and frequency of the power grid fluctuate within the normal range, avoiding the instability of the power grid caused by excessive frequency and voltage deviations. The real-time monitoring module can dynamically update the management strategy, enhance the flexibility and adaptability of the system, especially in the case of large load fluctuations, and can make adjustments quickly.

[0060] The real-time monitoring module predicts the fluctuation of electric vehicle charging load based on real-time monitoring data, including:

[0061] Based on the historical charging load data, the support vector machine is used to find an optimal hyperplane to divide these data into different categories. The support vector machine finds the optimal hyperplane that separates these data, specifically:

[0062] When satisfied Then i (w·x i +b)≥1,i=1,2,…,

[0063] That is, if for every data point x i , its classification label y i Satisfy i (w·x i+b)≥1, indicating that the data point is correctly classified and is outside the correct category boundary;

[0064] If any data point x i Not satisfied with y i (w·x i +b)≥1, the point is misclassified or falls on the boundary of the hyperplane, and the normal vector w and offset b of the hyperplane will continue to be optimized so that all points meet the conditions;

[0065] Among them, w represents the normal vector of the hyperplane, b represents the offset, i represents the index variable, and ||w|| represents the Euclidean norm of the vector w;

[0066] The final classification function is f(x), which is used to determine the category of load fluctuation. If the result of function f(x) is +1, it means that the current charging load is within the normal range. If f(x) is -1, it means that an abnormal load fluctuation is detected.

[0067] This embodiment uses a support vector machine to classify the historical load data of electric vehicle charging stations, with the goal of predicting future charging load fluctuations. The support vector machine divides different load data points into normal and abnormal categories by finding an optimal hyperplane. The goal of the support vector machine model is to maximize the boundary of the classified data, that is, to find the best normal vector and offset by optimizing the function. In actual operation, the load data is classified by the trained support vector machine model to obtain a classification function. When the value of the function is +1, the system determines that the current charging load is in the normal range; if the function result is -1, the system believes that the charging load fluctuation is abnormal and scheduling or adjustment measures may be required. This process relies on the analysis of historical data and compares the current data with the past pattern to detect potential load fluctuation anomalies in advance, thereby improving the system's responsiveness and stability. Through the application of the support vector machine algorithm, more accurate predictions can be made for electric vehicle charging load fluctuations to avoid grid instability caused by excessive fluctuations. The classification function can determine in real time whether the charging load is in a normal state, thereby timely discovering abnormal load fluctuations and preventing load overload or other grid risks. By automatically detecting and classifying the load state, the system can respond to sudden load fluctuations more quickly, reduce safety hazards caused by inaccurate predictions, and improve the operational reliability of the power grid. Based on the prediction and classification of machine learning algorithms, the new energy station management system can automatically respond to complex load changes without human intervention, thereby improving the overall system intelligence level.

[0068] The dynamic adjustment module dynamically adjusts the voltage and frequency of the distribution network and also includes:

[0069] The mean regression formula for voltage is:

[0070] V t+1 =V t +κ(μ V -V t ),

[0071] Among them, V t+1 represents the voltage value at the next moment, κ represents the regression rate constant, V t Indicates the voltage value at the current moment, μ V Indicates the mean voltage.

[0072] In this embodiment, the dynamic adjustment module dynamically adjusts the voltage value of the power grid through the voltage mean regression formula to ensure the stability of the voltage and maintain it near the ideal value. The formula calculates the voltage value V at the next moment based on the deviation between the current voltage value and the ideal voltage mean. t+1 , where the regression rate constant κ determines the speed at which the voltage returns to the ideal mean. V When the dynamic adjustment module calculates the voltage according to the formula, it will gradually restore the voltage to the ideal mean value, so as to maintain the stability of the grid voltage. If the regression speed constant κ value is large, the voltage will return faster, which is suitable for a rapidly changing load environment; while a smaller regression speed constant κ value is suitable for a relatively stable grid operation environment to prevent excessive voltage fluctuations from causing system instability. Through this regression formula, the system can automatically adjust the grid voltage to gradually return to the ideal voltage mean value μ V , to prevent the voltage from being too high or too low and affecting the normal operation of the power grid. By adjusting the regression speed constant κ, the speed of voltage adjustment can be controlled, so that dynamic adjustment can be flexibly performed according to different application scenarios. This formula can quickly adjust the voltage value when the load fluctuates, avoiding the problem of excessive voltage deviation during the operation of the power grid, ensuring the long-term stability and reliability of the power grid, and maintaining the voltage at the mean value μ V It can avoid the power loss caused by excessive voltage deviation nearby and improve the overall operation efficiency of the power grid.

[0073] The real-time monitoring module includes:

[0074] The high-dimensional features of the power grid state are extracted based on the deep belief network, and the energy function is used to characterize the correlation between input and output, specifically:

[0075]

[0076] Where E(v, h) represents the energy function, v i represents the input unit of the visible layer, h j represents the hidden layer unit, b i represents the bias parameter of the visible layer, c jrepresents the bias parameter of the hidden layer, and i and j both represent index variables.

[0077] In this embodiment, the real-time monitoring module extracts high-dimensional features of the power grid state by using a deep belief network. The model maps the input data of the power grid (such as voltage, current, frequency, etc.) to a high-dimensional hidden layer to extract more representative information. Specifically, an energy function is used to measure the relationship between the visible layer and the hidden layer:

[0078] The visible layer represents the monitoring data of the input power grid, such as voltage value, frequency, etc. The hidden layer is used to capture deeper features, which may be abstract representations of complex patterns of the power grid state. The bias parameter is used to adjust the flexibility of the model to ensure that the model can adapt to different power grid states. The weight parameter represents the connection strength between the visible layer and the hidden layer units. Through the training of these weights, the model can effectively extract high-dimensional features of the power grid state from the input data. The whole process optimizes the energy function through training and iteration to minimize the error between the input data and the model prediction results. Ultimately, the deep belief network can use these high-dimensional features to accurately monitor and analyze the power grid status in real time. By extracting the high-dimensional features of the power grid through the deep belief network, the system can capture complex patterns that are difficult to identify with traditional monitoring methods, providing more accurate and comprehensive power grid status monitoring. The use of deep learning models to automatically extract power grid features reduces reliance on artificial feature engineering and improves system efficiency. Especially when faced with large-scale power grid data, the deep belief network can quickly identify abnormal conditions in the power grid status through high-dimensional features, helping the system to make timely adjustments to avoid power grid failures or damage. Through the use of deep learning models, the system can more accurately predict power grid load fluctuations and other important power grid operating parameters, improving the management and optimization capabilities of the power grid.

[0079] The real-time monitoring module also includes collecting historical charging data and current charging demand of each charging station, analyzing these data, extracting charging behavior patterns, predicting future charging load fluctuations, and generating charging load fluctuation warnings based on the prediction results.

[0080] In this embodiment, the real-time monitoring module extracts the user's charging behavior pattern by comprehensively analyzing the historical charging data and current charging demand of each charging station. The collection and analysis of these data helps to predict future charging load fluctuations. The historical charging data reflects the changes in charging demand in different time periods. The system can identify the user's charging behavior pattern based on these data. For example, charging peaks in different time periods, common charging amount distribution, etc., and the current charging demand provides real-time load information. Based on this information and historical patterns, the system predicts future charging load fluctuations and generates load fluctuation warnings to help managers make charging scheduling plans in advance to cope with possible load peaks. By predicting future charging load fluctuations, the system can issue warnings in advance so that charging stations can take countermeasures to avoid grid instability or equipment overload caused by excessive load. By analyzing charging behavior patterns, the system can optimize scheduling plans, reasonably allocate charging resources, and reduce pressure during peak hours. Behavior pattern analysis based on historical data can help new energy stations manage charging resources more efficiently and improve the operational efficiency of stations. Through the combination of real-time data and historical data, the system can dynamically adjust the prediction model to ensure the accuracy and timeliness of the prediction results.

[0081] Collecting the historical charging data and current charging demand of each charging station specifically includes obtaining the historical charging data of each charging station, including the daily charging time and charging amount, and real-time collection of the current charging demand of each charging station, including the number of ongoing charging times and the number of scheduled charging times. Based on the historical data and current demand, the charging demand intensity of each charging station is calculated. When the calculated demand intensity I exceeds the preset threshold T, a demand peak warning is issued. The formula is: I>T, where I represents the charging demand intensity and T represents the preset threshold.

[0082] In this embodiment, the system calculates the charging demand intensity I by collecting the historical charging data and current charging demand of each charging station.

[0083] Historical data includes daily charging time and charging amount, which are used to analyze users' charging behavior patterns in different time periods;

[0084] Real-time data includes the number of charges currently in progress and the number of scheduled charges, which can provide current load information.

[0085] The system combines historical data with real-time data to calculate the demand intensity I of each charging station. When I exceeds the set threshold T, the system automatically issues a demand peak warning to help managers make advance scheduling and response.

[0086] By calculating the charging demand intensity I, the system can more accurately predict the load changes of each charging station and reduce the pressure on the power grid. When the demand intensity exceeds the preset threshold, the system will automatically issue an early warning to prompt managers to take corresponding measures to reduce the risk of overload during peak load periods. The system calculates through a combination of historical data and real-time data, and can more reasonably allocate charging resources and improve charging efficiency.

[0087] Based on historical data and current demand, the charging demand intensity of each charging station is calculated, including statistics on the average daily charging amount of each charging station in the past week, and statistics on the total charging amount currently being charged at each charging station. The current demand intensity is calculated based on these two. The formula is: I=C+(P / D), where I represents the charging demand intensity, C represents the average daily charging amount, P represents the current charging amount, and D represents the remaining time of the day. The calculation results are stored in the database.

[0088] This implementation further refines the calculation process of the charging demand intensity I. First, the average daily charging amount C of each charging station in the past week is counted, which is an average value obtained by analyzing historical data. Then the system counts the total charging amount P currently being charged in real time, and calculates it through the remaining time D of the day. The real-time charging demand is combined with the historical data to obtain the current demand intensity I. When the demand intensity exceeds the preset threshold, the system issues an early warning to remind managers to adjust the charging schedule. By combining historical and real-time data, the system can more accurately calculate the demand intensity of the charging station and improve the accuracy of load forecasting. By introducing the factor of the remaining time D of the day, the system can dynamically adjust the demand intensity calculation to adapt to changes in charging demand in different time periods. The calculation results are stored in the database, which is convenient for managers to conduct long-term data tracking and analysis, which helps to optimize long-term charging management strategies.

[0089] A new energy station management method based on a smart grid is also provided, using the new energy station management system based on a smart grid, the method comprising:

[0090] S1. Based on real-time monitoring data, predict the charging load fluctuation of electric vehicles. Based on the historical charging load data, find an optimal hyperplane through support vector machine, divide these data into different categories, and use classification function to judge the category of load fluctuation. If the function result is +1, it means that the current charging load is in the normal range. If it is -1, it means that the load fluctuation is abnormal.

[0091] S2. Adjust the output of new energy power generation equipment according to the predicted load fluctuations. According to the current action, the state of the power grid is transferred to a new state, and a reward is obtained after taking the action.

[0092] S3. Optimize the charging scheduling in the charging station to avoid overload caused by simultaneous peak charging. Define a fitness function to measure the pros and cons of each charging scheduling plan. Select a better charging scheduling plan based on the fitness function, and generate a new scheduling plan through crossover operations. Through multiple iterations, find the optimal charging scheduling strategy to avoid peak overload.

[0093] S4. Dynamically adjust the voltage and frequency of the distribution network to ensure stable operation, record the real-time values ​​of the voltage and frequency of the power grid, monitor the deviation of voltage and frequency in real time according to the ideal voltage and frequency averages of the power grid operation, set a threshold, and when the deviation exceeds the threshold, adjust the output and load distribution of the power generation equipment to gradually adjust the voltage and frequency back to the average value;

[0094] S5. Monitor the status of the distribution network in real time and extract high-dimensional features of the power grid status based on the deep belief network. Specifically, the features of the visible layer input units and hidden layer units are extracted through the energy function. The extracted features are used to monitor the power grid status in real time and dynamically update the management strategy to prevent overload.

[0095] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A new energy station management system based on smart grid, characterized in that: The system comprises: A prediction module, used to predict the fluctuation of electric vehicle charging load based on real-time monitoring data; The dispatching module is used to adjust the output of renewable energy power generation equipment according to the predicted load fluctuations; The optimization module is used to optimize the charging scheduling in the charging station to avoid overload caused by peak charging at the same time; the optimization module defines a fitness function f(p i ), measure the pros and cons of each scheduling scheme, and calculate the fitness function f(p i ) Select a scheduling plan and exchange scheduling plan information, generate a new scheduling plan, and iterate to find the optimal charging scheduling strategy to avoid peak overload; A dynamic adjustment module, used to dynamically adjust the voltage and frequency of the power distribution network; The dynamic adjustment module records the real-time values ​​of the voltage and frequency of the power grid V t and F t According to the ideal voltage mean μV and ideal frequency mean μF of the power grid operation, the deviation between voltage and frequency ΔV=V is monitored in real time. t -μV and ΔF = F t -μF, set a threshold value. If the deviation exceeds the threshold value, adjust the power generation equipment and load distribution in the power grid to gradually increase the voltage V t and frequency F t Adjust back to the mean μV and μF; The real-time monitoring module is used to monitor the status of the power distribution network in real time and dynamically update the management strategy.

2. According to claim 1, a new energy station management system based on smart grid is characterized by: The real-time monitoring module predicts the fluctuation of charging load of electric vehicles based on the real-time monitoring data, including: Based on the historical charging load data, the support vector machine is used to find an optimal hyperplane to divide these data into different categories. The support vector machine finds the optimal hyperplane that separates these data, specifically: When satisfied Then y i (w·x i +b)≥1, i=1,2,……, That is, if for every data point x i , its classification label y i Satisfy i (w·x i +b)≥1, indicating that the data point is correctly classified and is outside the correct category boundary; If any data point x i Not satisfied with y i (w·x i +b)≥1, the point is misclassified or falls on the boundary of the hyperplane, and the normal vector w and offset b of the hyperplane will continue to be optimized so that all points meet the conditions; Where w represents the normal vector of the hyperplane, b represents the offset, i represents the index variable, and ||w|| represents the Euclidean norm of the vector W; The final classification function is f(x), which is used to determine the category of load fluctuation. If the result of function f(x) is +1, it means that the current charging load is within the normal range. If f(x) is -1, it means that an abnormal load fluctuation is detected.

3. According to the smart grid-based new energy station management system of claim 1, it is characterized by: The dynamic adjustment module dynamically adjusts the voltage and frequency of the power distribution network and further includes: The mean regression formula for voltage is: V t+1 =V t +k(μV-V t ), Among them, V t+1 represents the voltage value at the next moment, k represents the regression rate constant, V t It represents the voltage value at the current moment, and μV represents the ideal voltage mean value.

4. According to claim 1, a new energy station management system based on smart grid is characterized by: The real-time monitoring module includes: The high-dimensional features of the power grid state are extracted based on the deep belief network, and the energy function is used to characterize the correlation between input and output, specifically: Where E(v, h) represents the energy function, v i represents the input unit of the visible layer, h j represents the hidden layer unit, b i represents the bias parameter of the visible layer, c j represents the bias parameter of the hidden layer, and i and j both represent index variables.

5. According to claim 1, a new energy station management system based on smart grid is characterized by: The scheduling module can transfer the state of the power grid to a new state according to the current action, and obtain a reward after taking a certain action.

6. According to claim 1, a new energy station management system based on smart grid is characterized by: The optimization module exchanges the information of the scheduling scheme and generates a new scheduling scheme. The specific process is as follows: p new =αp i +(1-α)p j , Among them, p new represents the new scheduling scheme, α represents the crossover ratio, and p i and p j Indicates two different scheduling schemes.

7. According to claim 1, a new energy station management system based on smart grid is characterized by: The real-time monitoring module also includes collecting historical charging data and current charging demand of each charging station, analyzing these data, extracting charging behavior patterns, predicting future charging load fluctuations, and generating charging load fluctuation warnings based on the prediction results.

8. According to claim 7, a new energy station management system based on smart grid is characterized in that: The collection of historical charging data and current charging demand of each charging station specifically includes obtaining historical charging data of each charging station, including daily charging time and charging amount, real-time collection of current charging demand of each charging station, including the number of ongoing charging times and the number of reserved charging times, and calculating the charging demand intensity of each charging station based on historical data and current demand. When the calculated demand intensity I exceeds a preset threshold value T, a demand peak warning is issued, and the formula is: I>T, where I represents the charging demand intensity and T represents the preset threshold value.

9. A new energy station management system based on smart grid according to claim 8, characterized in that: The charging demand intensity of each charging station is calculated based on historical data and current demand, including statistics on the average daily charging amount of each charging station in the past week, and statistics on the total charging amount currently being charged at each charging station. The current demand intensity is calculated based on these two, and the formula is: I=C+(P / D), where I represents the charging demand intensity, C represents the average daily charging amount, P represents the current charging amount, and D represents the remaining time of the day. The calculation results are stored in the database.

10. A new energy station management method based on smart grid, using the new energy station management system based on smart grid according to any one of claims 1 to 9, characterized in that: The method comprises: S1. Based on real-time monitoring data, predict the charging load fluctuation of electric vehicles. Based on the historical charging load data, find an optimal hyperplane through support vector machine, divide these data into different categories, and use classification function to judge the category of load fluctuation. If the function result is +1, it means that the current charging load is in the normal range. If it is -1, it means that the load fluctuation is abnormal. S2. Adjust the output of new energy power generation equipment according to the predicted load fluctuations. According to the current action, the state of the power grid is transferred to a new state, and a reward is obtained after taking the action. S3. Optimize the charging scheduling in the charging station to avoid overload caused by simultaneous peak charging. Define a fitness function to measure the pros and cons of each charging scheduling plan. Select a better charging scheduling plan based on the fitness function, and generate a new scheduling plan through crossover operations. Through multiple iterations, find the optimal charging scheduling strategy to avoid peak overload. S4. Dynamically adjust the voltage and frequency of the distribution network to ensure stable operation, record the real-time values ​​of the voltage and frequency of the power grid, monitor the deviation of voltage and frequency in real time according to the ideal voltage and frequency averages of the power grid operation, set a threshold, and when the deviation exceeds the threshold, adjust the output and load distribution of the power generation equipment to gradually adjust the voltage and frequency back to the average value; S5. Monitor the status of the distribution network in real time and extract high-dimensional features of the power grid status based on the deep belief network. Specifically, the features of the visible layer input units and hidden layer units are extracted through the energy function. The extracted features are used to monitor the power grid status in real time and dynamically update the management strategy to prevent overload.

Images