Micro-grid power optimization scheduling method and system

Abstract

The invention discloses a micro-grid power optimization scheduling method and system, and relates to the technical field of power systems, and the method comprises the steps: collecting power related data, building a micro-grid related data set S, and predicting the charging demands through a linear regression algorithm, thereby enabling the total charging demand C (t + k) of a charging station at a future time point t + k to be reasonably estimated, and achieving the optimal scheduling of the micro-grid power. Therefore, excessive deviation of the load of the charging station is effectively avoided, then, the load condition is monitored in real time and early warning is given out in time to ensure that the load of the charging station does not exceed the bearing capacity of the charging station, the overload risk of a power grid is avoided, in addition, the target function f (t) is optimized through a gradient descent algorithm, and the optimal charging power Popt of the charging station is dynamically adjusted. According to the method, charging requirements are accurately met, the situation that power fluctuation is too large and influences power grid stability is avoided, and finally, a scheduling scheme is adjusted in time in combination with real-time monitoring data so as to cope with load changes and optimize power grid resource configuration.

Description

Technical Field

[0001] The present invention relates to the technical field of power systems, and particularly to a method and system for optimizing power dispatching of a microgrid. Background Art

[0002] In the microgrid of a city, there are multiple electric vehicle (EV) charging stations distributed at various transportation hubs in the urban area. With the increase in traffic flow, the charging demand of electric vehicles rises sharply during peak hours. If the load of the charging stations is not regulated at this time, it will cause pressure on the microgrid, possibly leading to grid overload, voltage fluctuations, or even the shutdown of the charging stations. To avoid this situation, the system needs to perform intelligent dispatching by combining traffic flow and charging demand to ensure the stability of the grid and reduce power waste.

[0003] In a Chinese invention patent with the application publication number CN114069858B, a method for optimizing power dispatching of a power system is disclosed, which specifically relates to the technical field of switch control. It includes: a dispatching terminal, a substation, a distribution station, a transfer line, a central unit, an RTU remote measurement and control unit, a data exchange platform, a network data and telephone access terminal, a data exchange platform, a data fast channel, a local power grid rescue team. The dispatching general terminal is electrically connected to the data exchange platform through a wire, and the data exchange platform is signal-connected to the RTU remote measurement and control unit. The beneficial effect of this invention is that the RTU remote measurement and control unit sums up the fault information collected from the distribution station and the substation, and then transmits it to the data fast channel, and directly connects to the WEB server through the data fast channel. Then, power dispatching is performed through the dispatching console to ensure that if a power outage occurs during the normal operation of the project, the dispatching general terminal can respond quickly to ensure the project progress.

[0004] The above method can test the fault information of grid equipment and respond quickly when problems occur in the transfer line, ensuring the progress of the project. However, in addition to this, in the existing power dispatching methods, power dispatching is usually carried out after a fault occurs to repair the fault.

[0005] However, although this method can repair the faults of the grid, it will still cause certain losses and bring inconvenience to the grid users.

[0006] Therefore, the present invention provides a method and system for optimizing power dispatching of a microgrid. Summary of the Invention

[0007] (1) Technical Problems to be Solved

[0008] In view of the deficiencies of the prior art, the present invention provides a method and system for optimizing the power dispatching of a microgrid. A relevant data set S of the microgrid is established, providing accurate basic data for subsequent prediction and dispatching. The linear regression algorithm is used to accurately predict the charging demand, effectively avoiding excessive deviation of the charging station load. Then, the load situation is monitored in real time and early warnings are issued in a timely manner to ensure that the load of the charging station does not exceed its carrying capacity, avoiding the risk of grid overload. Moreover, the objective function f(t) is optimized by the gradient descent algorithm to dynamically adjust the optimal charging power of each charging station, ensuring the accurate satisfaction of the charging demand and avoiding excessive power fluctuations that affect the grid stability. Finally, combined with the real-time monitoring data, it can respond to load changes and optimize the grid resource allocation, maintain the grid load balance, reduce the overload risk, achieve a more efficient and stable power dispatching, and solve the problems in the above-mentioned background technology.

[0009] To achieve the above objectives, the present invention is realized through the following technical solutions: A method and system for optimizing the power dispatching of a microgrid, including a data acquisition module, a charging demand prediction module, a load assessment module, a load dispatching optimization module, and an intelligent control module;

[0010] The data acquisition module is used to collect power-related data in the microgrid coverage area, preprocess the collected power-related data, and construct a relevant data set S of the microgrid based on the preprocessed power-related data;

[0011] The charging demand prediction module is used to construct a charging demand prediction model using the linear regression algorithm, and obtain the total charging demand C(t + k) of the charging station at the future time point t + k based on the relevant data set S of the microgrid;

[0012] The load assessment module is used to obtain the charging station load score pf based on the relevant data set S of the microgrid and the total charging demand C(t + k), preset the charging station load threshold yz, compare and analyze the charging station load threshold yz and the charging station load score pf, evaluate the load capacity of the charging station in the future period of time. If the charging station load score pf is greater than or equal to the charging station load threshold yz, the alarm system is triggered for power dispatching optimization;

[0013] The load dispatching optimization module is used to construct the objective function f(t), set up constraint conditions, and optimize the objective function f(t) using the gradient descent algorithm to obtain the optimal charging power P of the charging station opt ;

[0014] The intelligent control module is used to construct a power optimization dispatching plan based on the optimal charging power P of the charging station opt , monitor the charging station power data in real time, evaluate the power optimization dispatching effect, and give feedback.

[0015] Preferably, the data acquisition module is used to connect to the charging station intelligent management system and the traffic flow monitoring system using an open API interface, and collect power-related data in the microgrid coverage area in real time, including the maximum output power P of the charging station max of the charging station, the maximum charging load C max of the charging station, the load E grid of the charging station, and the traffic flow V traffic , and preprocess the collected power-related data, including denoising, data cleaning, and data standardization;

[0016] Based on the preprocessed power-related data, construct a microgrid-related data set S.

[0017] Preferably, the charging demand prediction module is used to construct a charging demand prediction model using a linear regression algorithm, collect historical power-related data of the microgrid, and divide the historical power-related data into a training set and a validation set. Among them, the specific form of the charging demand prediction model is:

[0018] C(t + k)=β0 + β1·C(t)+β2·V traffic (t);

[0019] In the formula, C(t + k) represents the total charging demand of the charging station at the future time point t + k, C(t) represents the total charging demand of the charging station at the current time point t, V traffic (t) represents the traffic flow at the current time point t, β0 represents the bias term, and β1 and β2 represent the regression coefficients;

[0020] Input the training set into the charging demand prediction model for training the charging demand prediction model, and use the least squares method to solve the regression coefficients β1 and β2 of the charging demand prediction model, and use the validation set to verify the charging demand prediction model and optimize the parameters of the charging demand prediction model;

[0021] Based on the microgrid-related data set S, input the microgrid-related data set S into the trained charging demand prediction model to obtain the total charging demand C(t + k) of the charging station at the future time point t + k.

[0022] Preferably, the load evaluation module includes a data analysis unit and an evaluation unit;

[0023] The data analysis unit is used to obtain the charging station load score pf based on the microgrid-related data set S and the total charging demand C(t + k) of the charging station at the future time point t + k. The acquisition method of the charging station load score pf is:

[0024]

[0025] Wherein, C(t + k) represents the total charging demand of the charging station at the future time point t + k, and C max represents the maximum charging load of the charging station, n represents the total number of future time points predicted by the charging demand prediction model, and k = [1, 2, 3,..., n].

[0026] Preferably, the evaluation unit is used to preset a charging station load threshold yz, and compare and analyze the charging station load score pf and the charging station load threshold yz to evaluate the load capacity of the charging station in the future for a period of time. The specific evaluation content is as follows:

[0027] If the charging station load score pf is greater than or equal to the charging station load threshold yz, that is, pf ≥ yz, it is determined that the load capacity of the charging station is in the abnormal range in the future for a period of time, triggering the automatic alarm system, generating an alarm message, and sending the alarm message to the transfer dispatching management personnel to enter the load dispatching optimization module;

[0028] If the charging station load score pf is less than the charging station load threshold yz, that is, pf < yz, it is determined that the load capacity of the charging station is in the normal range in the future for a period of time and no processing is required.

[0029] Preferably, the load dispatching optimization module includes an objective function construction unit and a power dispatching optimization unit;

[0030] The objective function construction unit is used to construct an objective function f(t) and set constraint conditions. Among them, the specific expression form of the objective function f(t) is:

[0031]

[0032] Wherein, P i (t + k) represents the charging power of the i-th charging station at the future time point t + k, and C i (t + k) represents the total charging demand of the i-th charging station at the future time point t + k obtained by the charging demand prediction model, T represents the charging dispatching period, represents the demand power of the i-th charging station at the future time point t + k, λ represents the smoothing coefficient, and C(t + k - 1) represents the total charging demand of the i-th charging station at the future time point t + k - 1 obtained by the charging demand prediction model, represents the demand power of the i-th charging station at the future time point t + k - 1, and N is the total number of charging stations, i = [1, 2, 3,..., N];

[0033] The constraint conditions are:

[0034] P i (t + k) ≤ P i，max ;

[0035] Wherein, P i，max represents the maximum output power of the i-th charging station.

[0036] Preferably, the power dispatch optimization unit is used to optimize the objective function f(t) using the gradient descent method to obtain the optimal charging power of the charging station. The specific optimization includes initializing parameters, calculating the gradient of the objective function, and updating the charging power;

[0037] The initialization of parameters means setting the output power P i (t) of the i-th charging station at the current time point t as the initial value and setting the learning rate α;

[0038] The calculation of the gradient of the objective function means obtaining the gradient of the objective function according to the objective function f(t) using the partial derivative calculation method The gradient of the objective function The specific obtaining method is:

[0039]

[0040] The update of the charging power means updating the charging power of the charging station using the gradient descent formula. The specific form of the gradient descent formula is:

[0041]

[0042] Wherein, represents the updated charging power of the i-th charging station at the future time point t + k, represents the charging power of the i-th charging station before update at the future time point t + k, α represents the learning rate, represents the gradient of the objective function;

[0043] Repeat the update of the charging power, preset the change threshold ∈ of the objective function, and obtain the change value Δf(t) of the objective function. The obtaining method of the change value Δf(t) of the objective function is:

[0044]

[0045] Compare and analyze the objective function change threshold ∈ and the objective function change value Δf(t). When the objective function change value Δf(t) is less than the objective function change threshold ∈, that is, Δf(t) < ∈, stop the update of the charging power and output the charging power at this time as the optimal charging power P opt .

[0046] Preferably, the intelligent regulation module includes a power regulation unit and an effect evaluation unit;

[0047] The power regulation unit is used to according to the optimal charging power P of the charging stationopt , construct an optimal power dispatch plan, send the optimal power dispatch plan to the microgrid control system, generate control instructions, and perform optimal power dispatch for the microgrid according to the control instructions, including power control and energy dispatch of the charging station, and continuously monitor the power output status of the charging station. Among them, the optimal power dispatch plan includes the optimal charging power of each charging station within the area covered by the microgrid.

[0048] Preferably, the effect evaluation unit is used to collect power-related data in real time after the optimal power dispatch, and combine the total charging demand C(t + k) at the future time point t + k obtained by the charging demand prediction model to calculate the demand satisfaction error DSE. The way to obtain the demand satisfaction error DSE is as follows:

[0049]

[0050] In the formula, represents the total charging amount of the i-th charging station at the future time point t + k, represents the total charging demand C(t + k) of the i-th charging station at the future time point t + k obtained by the charging demand prediction model;

[0051] If DSE is greater than or equal to zero, it is determined that the optimal power dispatch effect is normal, and the optimal power dispatch process and dispatch data are recorded. If DSE is less than zero, it is determined that the optimal power dispatch effect is abnormal. At this time, power-related data is collected and feedback is made, and the load dispatch optimization of the microgrid is carried out again.

[0052] A microgrid optimal power dispatch method includes the following steps:

[0053] Step 1: Collect power-related data in the area covered by the microgrid, preprocess the collected power-related data, and construct a microgrid-related data set S based on the preprocessed power-related data;

[0054] Step 2: Use the linear regression algorithm to construct a charging demand prediction model, and obtain the total charging demand C(t + k) of the charging station at the future time point t + k according to the microgrid-related data set S;

[0055] Step 3: According to the microgrid-related data set S and the total charging demand C(t + k), obtain the charging station load score pf, preset the charging station load threshold yz, compare and analyze the charging station load threshold yz and the charging station load score pf, and evaluate the load capacity of the charging station in the future period. If the charging station load score pf is greater than or equal to the charging station load threshold yz, trigger the alarm system and perform power dispatch optimization;

[0056] Step 4: Construct the objective function f(t), set up constraint conditions, and use the gradient descent algorithm to optimize the objective function f(t) to obtain the optimal charging power P of the charging station opt ;

[0057] Step 5: Based on the optimal charging power P of the charging station opt , construct a power optimization scheduling plan, monitor the power data of the charging station in real time, evaluate the effect of the power optimization scheduling, and give feedback.

[0058] The present invention provides a method and system for optimizing the power scheduling of a microgrid, having the following beneficial effects:

[0059] (1) Through the microgrid power optimization scheduling system, the system can collect power data, charging demand data, and traffic flow data in the microgrid in real time. Using an advanced charging demand prediction model and a load evaluation module, predict the total charging demand C(t + k) at a future time point t + k, and make a pre-judgment based on the charging station load score pf. If the charging station load score pf is greater than or equal to the charging station load threshold yz, the system can timely trigger an alarm and perform scheduling optimization. Through these steps, the system can avoid the overloading phenomenon of the charging station, improve the balance of the grid load and the utilization efficiency of resources at the same time. This process effectively reduces the waste of grid energy, optimizes the power distribution of the charging station, and improves the overall utilization efficiency of power resources.

[0060] (2) Through the real-time monitoring and feedback mechanism, combined with the objective function optimization and the gradient descent algorithm, dynamically optimize the charging power of the charging station, so as to achieve the stable operation of power scheduling. The system accurately schedules the power of each charging station based on real-time power-related data and demand prediction, and ensures that during peak load periods, the stability of the grid is hardly affected. At the same time, through the load scheduling optimization module, avoid the overloading of the charging station power, thereby effectively preventing the power instability or voltage fluctuation problems of the microgrid due to excessive load, and ensuring the stability and long-term reliable operation of the microgrid.

[0061] (3) Through the collaborative work of the data acquisition module and the intelligent control module, the system can quickly respond to the changes in the grid load and charging demand. With real-time data collection and dynamic charging demand prediction, the microgrid can self-adjust according to the actual demand and environmental factors. When the charging station load score pf of a certain charging station is greater than or equal to the charging station load threshold yz, the system will immediately trigger scheduling optimization and avoid overload by adjusting the charging power. In addition, the adaptive load scheduling of the system can flexibly handle different charging demands and grid load conditions, ensuring that the microgrid can work efficiently under various operating conditions. This flexibility enables the microgrid to quickly adapt to changing demands, improving the overall system response speed and scheduling efficiency. Description of the Drawings

[0062] Figure 1 This is a schematic diagram of the block diagram process of a microgrid power optimization and dispatch system according to the present invention.

[0063] Figure 2 This is a schematic diagram of the process of a microgrid power optimization and dispatch method according to the present invention. Specific embodiments

[0064] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0065] Embodiment 1

[0066] Please refer to Figure 1 , the present invention provides a microgrid power optimization and dispatch system, including a data acquisition module, a charging demand prediction module, a load evaluation module, a load dispatch optimization module, and an intelligent control module;

[0067] The data acquisition module is used to collect power-related data in the microgrid coverage area, preprocess the collected power-related data, and construct a microgrid-related data set S based on the preprocessed power-related data;

[0068] The charging demand prediction module is used to construct a charging demand prediction model using a linear regression algorithm, and obtain the total charging demand C(t + k) of the charging station at the future time point t + k according to the microgrid-related data set S;

[0069] The load evaluation module is used to obtain the charging station load score pf according to the microgrid-related data set S and the total charging demand C(t + k), preset the charging station load threshold yz, compare and analyze the charging station load threshold yz and the charging station load score pf, evaluate the load capacity of the charging station in the future period of time. If the charging station load score pf is greater than or equal to the charging station load threshold yz, the alarm system is triggered for power dispatch optimization;

[0070] The load dispatch optimization module is used to construct an objective function f(t), set up constraint conditions, and optimize the objective function f(t) using a gradient descent algorithm to obtain the optimal charging power P of the charging station opt ;

[0071] The intelligent control module is used to construct a power optimization and dispatch plan according to the optimal charging power P of the charging station opt , and real-time monitor the charging station power data, evaluate the power optimization and dispatch effect, and give feedback.

[0072] In the embodiment, through the collaborative work of the data acquisition module, the charging demand prediction module, the load assessment module, the load scheduling optimization module, and the intelligent regulation module, the scheduling efficiency, stability, and response ability of the microgrid are improved. The data acquisition module collects real-time power-related data within the microgrid area and preprocesses the power-related data to construct a complete set S of microgrid-related data, providing an accurate data basis for subsequent decision-making. The charging demand prediction module accurately predicts the charging demand in future periods through a linear regression algorithm, combining historical data and current data, helping the system make more scientific scheduling decisions. The load assessment module calculates the charging station load score pf and, by comparing it with a preset threshold, promptly discovers the risk of load overload and triggers an alarm mechanism to ensure that the charging station load does not exceed the safe range. The load scheduling optimization module constructs an optimization objective function f(t) and uses the gradient descent method for optimization to calculate the optimal charging power for each charging station, meeting the charging demand while avoiding grid overload. Finally, the intelligent regulation module formulates a charging station scheduling plan based on the optimal charging power of the charging station and monitors the scheduling effect in real time to ensure the execution effect of the power scheduling plan. Overall, this system improves the operating efficiency and safety of the power grid, ensures the efficient satisfaction of charging demands, optimizes the resource utilization of the microgrid, and enhances the flexibility and stability of the system.

[0073] Embodiment 2

[0074] Please refer to Figure 1 Specifically: The data acquisition module is used to connect to the charging station intelligent management system and the traffic flow monitoring system using an open API interface, and collect real-time power-related data in the area covered by the microgrid, including the maximum output power P of the charging station max , the maximum charging load C of the charging station max , the charging station load E grid and the traffic flow V traffic , and preprocess the collected power-related data, including denoising, data cleaning, and data standardization;

[0075] The maximum output power P of the charging station max , the maximum charging load C of the charging station max and the charging station load E grid are obtained through the charging station intelligent management system;

[0076] The traffic flow V traffic is obtained through the traffic flow monitoring system;

[0077] Based on the preprocessed power-related data, a set S of microgrid-related data is constructed.

[0078] In an embodiment, through the efficient connection of the open API interface, it is possible to collect power-related data within the microgrid coverage area in real time, including the maximum output power P of the charging station max , the maximum charging load C of the charging station max , the load E of the charging station grid and the traffic flow V traffic . This function enables the system to comprehensively grasp the real-time status of the power grid and the charging station, as well as the charging demand fluctuations caused by changes in traffic flow. This real-time data collection ability provides accurate basic data for subsequent charging demand prediction and load scheduling. During the data preprocessing process, denoising, data cleaning, and standardization operations ensure the high quality of the data, avoiding interference from abnormal data in subsequent analysis and decision-making. Through these processes, the microgrid-related data set S is constructed, containing comprehensive and accurate power-related information, providing a reliable input data source for modules such as charging demand prediction, load assessment, and optimal scheduling. This process improves the decision-making accuracy of the microgrid system, enhances the system's ability to respond to load fluctuations, helps achieve grid stability, efficiently meet charging demands, and reasonably allocate resources, thus laying a data support foundation for the efficient operation and optimal scheduling of the microgrid.

[0079] Embodiment 3

[0080] Please refer to Figure 1 , specifically: The charging demand prediction module is used to construct a charging demand prediction model using the linear regression algorithm, collect historical power-related data of the microgrid, and divide the historical power-related data into a training set and a validation set. Among them, the specific form of the charging demand prediction model is:

[0081] C(t + k) = β0 + β1·C(t) + β2·V traffic (t);

[0082] In the formula, C(t + k) represents the total charging demand of the charging station at the future time point t + k, C(t) represents the total charging demand of the charging station at the current time point t, V traffic (t) represents the traffic flow at the current time point t, β0 represents the bias term, and β1 and β2 represent the regression coefficients;

[0083] Input the training set into the charging demand prediction model, train the charging demand prediction model, and use the least squares method to solve the regression coefficients β1 and β2 of the charging demand prediction model, and use the validation set to verify the charging demand prediction model and optimize the parameters of the charging demand prediction model;

[0084] According to the microgrid-related data set S, input the microgrid-related data set S into the trained charging demand prediction model to obtain the total charging demand C(t + k) of the charging station at the future time point t + k.

[0085] In the embodiment, by applying the linear regression algorithm and using historical power-related data to construct a charging demand prediction model, the prediction accuracy of the microgrid for future charging demand is improved. Through the collection and analysis of the microgrid historical data, the system divides the data into a training set and a validation set to ensure the effectiveness of model training and validation. During the training process, the least squares method is used to optimize the regression coefficients, and the model parameters are continuously adjusted through the validation set, thereby obtaining a high-precision charging demand prediction model. This model can accurately predict the total charging demand C(t + k) of the charging station at the future time point t + k, providing a scientific basis for power dispatching and load distribution. By inputting the real-time microgrid-related data set S into the trained charging demand prediction model, the system can dynamically respond to the changes in charging demand, ensuring the real-time update and optimization of the charging station dispatching plan. This not only effectively avoids the problem of grid overload when the charging demand is too high but also can reasonably allocate resources during the low-demand period, improving the utilization efficiency of the charging station and optimizing the overall load balance of the grid. This dispatching optimization scheme based on accurate demand prediction ensures the high efficiency and stability of the microgrid operation and improves the energy use efficiency, with significant economic and environmental benefits.

[0086] Embodiment 4

[0087] Please refer to Figure 1 , specifically: the load evaluation module includes a data analysis unit and an evaluation unit;

[0088] The data analysis unit is used to obtain the charging station load score pf based on the microgrid-related data set S and the total charging demand C(t + k) of the charging station at the future time point t + k. The method for obtaining the charging station load score pf is as follows:

[0089]

[0090] In the formula, C(t + k) represents the total charging demand of the charging station at the future time point t + k, C max represents the maximum charging load of the charging station, n represents the total number of future time points predicted by the charging demand prediction model, and k = [1, 2, 3,..., n].

[0091] The evaluation unit is used to preset the charging station load threshold yz and compare and analyze the charging station load score pf and the charging station load threshold yz to evaluate the load capacity of the charging station within a future period of time. The specific evaluation content is as follows:

[0092] If the charging station load score pf is greater than or equal to the charging station load threshold yz, i.e., pf≥yz, it is determined that the load capacity of the charging station is in the abnormal range within a certain period of time in the future, triggering the automatic alarm system, generating an alarm message, and sending the alarm message to the dispatching management personnel for transfer, and entering the load dispatching optimization module;

[0093] If the charging station load score pf is less than the charging station load threshold yz, i.e., pf<yz, it is determined that the load capacity of the charging station is in the normal range within a certain period of time in the future, and no processing is required.

[0094] In the embodiment, through the implementation of the load assessment module, the system can effectively evaluate and monitor the load capacity of the charging station in real-time dynamic dispatching, ensuring the smoothness and stability of the microgrid operation. The data analysis unit calculates the charging station load score pf based on the relevant data set S of the microgrid and the total charging demand C(t + k) of the charging station at the future time point t + k, providing a scientific basis for load assessment. According to the charging station load score pf, the assessment unit can compare and analyze the charging station load score pf with the preset charging station load threshold yz to achieve intelligent monitoring of the charging station load capacity. When the charging station load score pf is greater than or equal to the charging station load threshold yz, the system automatically triggers the alarm mechanism to remind the dispatching management personnel to take corresponding optimization measures. This mechanism effectively prevents the charging station from overloading during high-demand periods, ensuring the stability of the power grid and the normal operation of the charging station. On the contrary, when the charging station load score pf is less than the charging station load threshold yz, the system determines that the load capacity of the charging station is in the normal range, thus avoiding unnecessary intervention and reducing the unnecessary burden on the system. Through this intelligent evaluation and control process, the system can respond to the load fluctuations of the microgrid in real-time, timely adjust the power distribution plan, improve the accuracy and efficiency of the charging station load dispatching, ultimately optimize the utilization rate of power grid resources, and reduce the operation risk of the power grid.

[0095] Embodiment 5

[0096] Please refer to Figure 1 , specifically: the load dispatching optimization module includes an objective function construction unit and a power dispatching optimization unit;

[0097] The objective function construction unit is used to construct the objective function f(t) and set up constraint conditions, where the specific form of the objective function f(t) is:

[0098]

[0099] In the formula, P i (t + k) represents the charging power of the i-th charging station at the future time point t + k, C i(t + k) represents the total charging demand of the i-th charging station at the future time point t + k obtained by the charging demand prediction model, T represents the charging scheduling period, represents the demand power of the i-th charging station at the future time point t + k, λ represents the smoothing coefficient, and C(t + k - 1) represents the total charging demand of the i-th charging station at the future time point t + k - 1 obtained by the charging demand prediction model, represents the demand power of the i-th charging station at the future time point t + k - 1, N is the total number of charging stations, and i = 1, 2, 3, …, N];

[0100] The charging scheduling period T is usually set to 1 hour, representing the time span of each scheduling and optimization, and the period length can also be adjusted according to requirements;

[0101] The smoothing coefficient λ is a parameter set by the system, which determines the degree of power smoothing. A larger λ value will result in a greater penalty for power fluctuations, and the system will pay more attention to power stability;

[0102] The measures the gap between the actual charging power and the demand power. The goal is to minimize this gap to ensure that the power of the charging station meets the demand as much as possible.

[0103] The is a penalty term that reflects the penalty for power changes. The purpose is to smooth the power fluctuations of the charging station and avoid excessive power changes between adjacent periods, thereby maintaining the stability of the power grid.

[0104] The constraint conditions are:

[0105] P i (t + k) ≤ P i，max ;

[0106] In the formula, P i，max represents the maximum output power of the i-th charging station.

[0107] The power dispatch optimization unit is used to optimize the objective function f(t) using the gradient descent method to obtain the optimal charging power of the charging station. The specific optimization includes parameter initialization, objective function gradient calculation, and charging power update;

[0108] The gradient descent method is an iterative optimization algorithm used to find the local minimum of the objective function f(t). Its basic idea is to calculate the gradient of the objective function f(t), that is, the partial derivative, and then update the parameters along the opposite direction of the gradient to gradually approach the minimum value of the objective function f(t);

[0109] The main benefit of using the gradient descent method to optimize the objective function f(t) in the microgrid power dispatch is:

[0110] Efficient calculation: The gradient descent method can quickly find the optimal solution by calculating the gradient, avoiding heavy global search;

[0111] Simple and easy to implement: The algorithm structure of the gradient descent method is simple and can be quickly implemented in a large-scale power dispatching system;

[0112] Strong adaptability: It can adapt to multi-variable optimization problems and provide real-time scheduling under the changing charging demands and grid loads;

[0113] The initialization parameter refers to setting the output power P i (t) of the i-th charging station at the current time point t as the initial value and setting the learning rate α;

[0114] The calculation of the gradient of the objective function refers to obtaining the gradient of the objective function using the partial derivative calculation method based on the objective function f(t); The gradient of the objective function The specific obtaining method is:

[0115]

[0116] The update of the charging power refers to updating the charging power of the charging station using the gradient descent formula, and the specific expression form of the gradient descent formula is:

[0117]

[0118] In the formula, represents the updated charging power of the i-th charging station at the future time point t + k, represents the charging power of the i-th charging station before update at the future time point t + k, α represents the learning rate, represents the gradient of the objective function;

[0119] The learning rate α is a parameter that controls the step size of the update. If the learning rate α is too small, the optimization process will be very slow and more iterations are required to find the optimal solution. If the learning rate α is too large, it may cause the objective function f(t) to skip the optimal solution during the optimization process, or even make the algorithm diverge, resulting in an inability to find a convergent solution. The learning rate can be adjusted through the Adam algorithm, combining the mean and root mean square value of the gradient.

[0120] Repeat the update of the charging power, preset the change threshold ∈ of the objective function, and obtain the change value Δf(t) of the objective function. The obtaining method of the change value Δf(t) of the objective function is:

[0121]

[0122] Compare and analyze the target function change threshold ∈ and the target function change value Δf(t). When the target function change value Δf(t) is less than the target function change threshold ∈, that is, when Δf(t) < ∈, stop updating the charging power and output the charging power at this time as the optimal charging power P of the charging station. opt .

[0123] In the embodiment, the charging power of the charging station is optimized by the gradient descent method to achieve reasonable load distribution and grid balance. This module first constructs an optimization target function f(t) based on the total charging demand C(t) and charging power demand of the charging station at the future time point t + k provided by the charging demand prediction model, combined with factors such as the smoothing coefficient and the scheduling period, and sets the constraint condition of the maximum power output of the charging station. The optimization of the target function f(t) will minimize the error between the actual power and the demand power of the charging station, while controlling the power fluctuation, so as to achieve stable load scheduling. By using the gradient descent method, the system can gradually adjust and optimize the output power of the charging station according to the current charging power to achieve the optimal charging power distribution. By continuously iterating and updating the charging power until the target function change value Δf(t) is less than the preset target function change threshold ∈, convergence is ensured, and finally the optimal charging power of the charging station is obtained. This optimization process improves the resource utilization efficiency of the charging station, while avoiding excessive fluctuations in the charging power and ensuring grid stability. The beneficial effect of this load scheduling optimization module is that it can accurately adjust the power output of each charging station, ensure grid load balance, avoid overload, and reasonably allocate grid resources, effectively improving the stability of the microgrid operation and the satisfaction of charging demand.

[0124] Embodiment 6

[0125] Please refer to Figure 1 , specifically: The intelligent control module includes a power control unit and an effect evaluation unit;

[0126] The power control unit is used to construct a power optimization scheduling plan based on the optimal charging power P of the charging station opt , and send the power optimization scheduling plan to the microgrid control system to generate control instructions, and perform power optimization scheduling on the microgrid according to the control instructions, including power control and energy scheduling of the charging station, and continuously monitor the power output status of the charging station. Among them, the power optimization scheduling plan includes the optimal charging power of each charging station in the area covered by the microgrid.

[0127] The effect evaluation unit is used to collect power-related data in real time after power optimization scheduling, and calculate the demand satisfaction error DSE in combination with the total charging demand C(t + k) at the future time point t + k obtained by the charging demand prediction model. The method for obtaining the demand satisfaction error DSE is:

[0128]

[0129] Wherein, represents the total charging amount of the i-th charging station at the future time point t + k, represents the total charging demand C(t + k) of the i-th charging station at the future time point t + k obtained by the charging demand prediction model;

[0130] If DSE is greater than or equal to zero, it is determined that the power optimization scheduling effect is normal, and the power optimization scheduling process and scheduling data are recorded. If DSE is less than zero, it is determined that the power optimization scheduling effect is normal. At this time, power-related data is collected and feedback is given, and the load scheduling optimization of the microgrid is re-performed.

[0131] In the embodiment, through the collaborative work of the power control unit and the effect evaluation unit, the intelligent level and efficiency of the microgrid power scheduling are improved. The power control unit constructs an accurate power optimization scheduling scheme according to the optimal charging power of the charging station, and converts the scheme into a control instruction to implement the scheduling through the microgrid control system. This process includes the real-time control of the charging station power and the reasonable scheduling of energy, which not only ensures the balance of power distribution, but also effectively prevents the grid load from overloading. At the same time, by continuously monitoring the power output of the charging station, the system can timely adjust the charging strategy, optimize the energy use, avoid unnecessary power waste, improve the response speed and reliability of the system. The effect evaluation unit further evaluates the power optimization scheduling effect by calculating the demand satisfaction error DSE to ensure that the power optimization scheduling can effectively meet the expected charging demand. If the power optimization scheduling meets the expectation, the system will record the scheduling process and data; if it is found that the load is not reasonably distributed, the system can give real-time feedback and make adjustments to ensure the continuous optimization of the power scheduling. This mechanism based on real-time data feedback and self-adjustment improves the flexibility of the system and the ability to cope with emergencies, provides a more efficient, stable and sustainable power scheduling solution for the microgrid, and significantly improves the intelligent level of energy management.

[0132] Embodiment 7

[0133] Please refer to Figure 2 , specifically: A microgrid power optimization scheduling method includes the following steps.

[0134] Step 1: Collect power-related data in the microgrid coverage area, preprocess the collected power-related data, and construct a microgrid-related data set S based on the preprocessed power-related data.

[0135] Step 2: Use the linear regression algorithm to construct a charging demand prediction model, and obtain the total charging demand C(t + k) of the charging station at the future time point t + k according to the microgrid-related data set S.

[0136] Step 3: Based on the microgrid - related data set S and the total charging demand C(t + k), obtain the charging station load score pf, and preset the charging station load threshold yz. Compare and analyze the charging station load threshold yz and the charging station load score pf to evaluate the load capacity of the charging station in the next period of time. If the charging station load score pf is greater than or equal to the charging station load threshold yz, trigger the alarm system and perform power dispatch optimization;

[0137] Step 4: Construct the objective function f(t), set up constraints, and use the gradient - descent algorithm to optimize the objective function f(t) to obtain the optimal charging power P of the charging station opt ;

[0138] Step 5: Based on the optimal charging power P of the charging station opt , construct a power optimization dispatch plan, and monitor the charging station power data in real - time, evaluate the power optimization dispatch effect, and give feedback.

[0139] In the embodiment, by collecting and pre - processing power - related data, a microgrid - related data set S is established, providing accurate basic data for subsequent prediction and dispatch. The linear regression algorithm is used to accurately predict the charging demand, so that the total charging demand C(t + k) at the future time point t + k is reasonably estimated, effectively avoiding excessive deviation of the charging station load. Then, by combining the total charging demand C(t + k) and the charging station load threshold yz, the load situation is monitored in real - time and early warnings are issued in a timely manner to ensure that the load of the charging station does not exceed its carrying capacity and avoid the risk of grid overload. In addition, by using the gradient - descent algorithm to optimize the objective function f(t), the system can dynamically adjust the optimal charging power of each charging station to ensure accurate satisfaction of the charging demand and avoid excessive power fluctuations affecting grid stability. Finally, combined with real - time monitoring data, the system can timely adjust the dispatch plan to cope with load changes and optimize the grid resource allocation. In summary, this dispatch method effectively improves the grid load balance, reduces the overload risk, and increases the response speed and flexibility of the microgrid, achieving more efficient and stable power dispatch.

[0140] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principle and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A microgrid power optimization dispatch system, characterized in that: It includes a data acquisition module, a charging demand prediction module, a load assessment module, a load scheduling optimization module, and an intelligent regulation module; The data acquisition module is used to collect power-related data in the microgrid coverage area, preprocess the collected power-related data, and construct a microgrid-related data set S based on the preprocessed power-related data; The charging demand prediction module is used to construct a charging demand prediction model using a linear regression algorithm, and based on the microgrid-related data set S, obtain the total charging demand C(t + k) of the charging station at the future time point t + k; The load assessment module is used to obtain the charging station load score pf based on the microgrid-related data set S and the total charging demand C(t + k), preset a charging station load threshold yz, compare and analyze the charging station load threshold yz and the charging station load score pf, and evaluate the load capacity of the charging station in the future for a period of time. If the charging station load score pf is greater than or equal to the charging station load threshold yz, the alarm system is triggered for power scheduling optimization; The load scheduling optimization module is used to construct the objective function f(t), set up constraint conditions, and optimize the objective function f(t) using the gradient descent algorithm to obtain the optimal charging power P of the charging station opt ; The intelligent control module is used to adjust the optimal charging power P of the charging station according to the opt , build a power optimization scheduling plan, monitor the power data of charging stations in real time, evaluate the power optimization scheduling effect, and provide feedback.

2. The microgrid power optimization scheduling system according to claim 1, wherein: The data acquisition module is used to connect to the charging station intelligent management system and the traffic flow monitoring system using an open API interface, and collect power-related data in the microgrid coverage area in real time, including the maximum output power P of the charging station max , the maximum charging load C of the charging station max , the load E of the charging station grid and the traffic flow V traffic , and preprocess the collected power-related data, including denoising, data cleaning, and data standardization; Based on the preprocessed power-related data, construct a microgrid-related data set S.

3. The microgrid power optimization scheduling system according to claim 2, characterized in that: The charging demand prediction module is used to construct a charging demand prediction model using a linear regression algorithm, collect the historical power-related data of the microgrid, and divide the historical power-related data into a training set and a validation set. Among them, the specific form of the charging demand prediction model is: C(t + k) = β0 + β1·C(t) + β2·V traffic (t); Where, C(t + k) represents the total charging demand of the charging station at the future time point t + k, C(t) represents the total charging demand of the charging station at the current time point t, V traffic (t) represents the traffic flow at the current time point t, β0 represents the bias term, and β1 and β2 represent the regression coefficients; Input the training set into the charging demand prediction model for training the charging demand prediction model, use the least squares method to solve the regression coefficients β1 and β2 of the charging demand prediction model, and use the validation set to verify the charging demand prediction model and optimize the parameters of the charging demand prediction model; Based on the microgrid-related data set S, input the microgrid-related data set S into the trained charging demand prediction model to obtain the total charging demand C(t + k) of the charging station at the future time point t + k.

4. A microgrid power optimization dispatch system according to claim 3, characterized in that: The load assessment module includes a data analysis unit and an assessment unit; The data analysis unit is used to obtain the charging station load score pf based on the microgrid-related data set S and the total charging demand C(t + k) of the charging station at the future time point t + k. The method for obtaining the charging station load score pf is: Wherein, C(t + k) represents the total charging demand of the charging station at the future time point t + k, and C max represents the maximum charging load of the charging station, n represents the total number of future time points predicted by the charging demand prediction model, and k = 1, 2, 3, …, n.

5. The microgrid power optimization scheduling system according to claim 4, characterized in that: The assessment unit is used to preset a charging station load threshold yz, and compare and analyze the charging station load score pf and the charging station load threshold yz to evaluate the load capacity of the charging station in the future for a period of time. The specific assessment content is as follows: If the charging station load score pf is greater than or equal to the charging station load threshold yz, that is, pf ≥ yz, it is determined that the load capacity of the charging station is in the abnormal range in the future for a period of time, trigger the automatic alarm system, generate an alarm message, and send the alarm message to the transfer dispatching management personnel to enter the load scheduling optimization module; If the charging station load score pf is less than the charging station load threshold yz, that is, pf < yz, it is determined that the load capacity of the charging station is in the normal range in the future for a period of time and no processing is required.

6. The microgrid power optimization dispatch system according to claim 5, characterized in that: The load scheduling optimization module includes an objective function construction unit and a power scheduling optimization unit; The objective function construction unit is used to construct the objective function f(t) and set up constraint conditions. Among them, the specific form of the objective function f(t) is as follows: Where P i (t + k) represents the charging power of the i-th charging station at the future time point t + k, C i (t + k) represents the total charging demand of the i-th charging station at the future time point t + k obtained by the charging demand prediction model, T represents the charging scheduling period, represents the demand power of the i-th charging station at the future time point t + k, λ represents the smoothing coefficient, C(t + k - 1) represents the total charging demand of the i-th charging station at the future time point t + k - 1 obtained by the charging demand prediction model, represents the demand power of the i-th charging station at the future time point t + k - 1, N is the total number of charging stations, i = [1, 2, 3,..., N]; The constraint conditions are: P i (t + k) ≤ P i，max ; Wherein, P i，max represents the maximum output power of the i-th charging station.

7. The microgrid power optimization dispatch system according to claim 6, characterized in that: The power dispatch optimization unit is used to optimize the objective function f(t) using the gradient descent method to obtain the optimal charging power of the charging station. The specific optimization includes initializing parameters, calculating the gradient of the objective function, and updating the charging power; The initialization parameter refers to setting the output power P i (t) of the i-th charging station at the current time point t as the initial value and setting the learning rate α; The calculation of the gradient of the objective function refers to obtaining the gradient of the objective function according to the objective function f(t) using the method of calculating partial derivatives. The gradient of the objective function The specific obtaining method is as follows: The charging power update refers to updating the charging power of the charging station using the gradient descent formula. The specific form of the gradient descent formula is as follows: In the formula, represents the updated charging power of the i-th charging station at the future time point t + k, represents the charging power of the i-th charging station before update at the future time point t + k, and α represents the learning rate. represents the gradient of the objective function; Repeat the charging power update, preset the objective function change threshold ∈, and obtain the objective function change value Δf(t). The way to obtain the objective function change value Δf(t) is: Compare and analyze the target function change threshold ε and the target function change value Δf(t). When the target function change value Δf(t) is less than the target function change threshold ε, that is, when Δf(t) < ε, stop updating the charging power and output the charging power at this time as the optimal charging power P of the charging station opt .

8. The microgrid power optimization scheduling system according to claim 7, characterized in that: The intelligent regulation module includes a power regulation unit and an effect evaluation unit; The power regulation unit is used to construct a power optimization scheduling plan according to the optimal charging power P of the charging station, and send the power optimization scheduling plan to the microgrid control system to generate control instructions, and perform power optimization scheduling on the microgrid according to the control instructions, including power control and energy scheduling of the charging station, and continuously monitor the power output status of the charging station. Among them, the power optimization scheduling plan includes the optimal charging power of each charging station in the area covered by the microgrid. opt ​ 9. The microgrid power optimization scheduling system according to claim 8, wherein: The effect evaluation unit is used to collect the power-related data after the power optimization dispatch in real time, and combine the total charging demand C(t + k) at the future time point t + k obtained by the charging demand prediction model to calculate the demand satisfaction error DSE. The way to obtain the demand satisfaction error DSE is: In the formula, represents the total charging amount of the i-th charging station at the future time point t + k, represents the total charging demand C(t + k) of the i-th charging station at the future time point t + k obtained by the charging demand prediction model; If DSE is greater than or equal to zero, it is determined that the power optimization dispatch effect is normal, and the power optimization dispatch process and dispatch data are recorded. If DSE is less than zero, it is determined that the power optimization dispatch effect is normal. At this time, collect the power-related data and give feedback, and re-optimize the load dispatch of the microgrid.

10. A microgrid power optimization dispatch method for implementing the microgrid power optimization dispatch system according to any one of claims 1 to 9 above, characterized in that: It includes the following steps Step 1: Collect the power-related data in the microgrid coverage area, preprocess the collected power-related data, and construct the microgrid-related data set S based on the preprocessed power-related data; Step 2: Use the linear regression algorithm to construct the charging demand prediction model, and obtain the total charging demand C(t + k) of the charging station at the future time point t + k based on the microgrid-related data set S; Step 3: Based on the microgrid-related data set S and the total charging demand C(t + k), obtain the charging station load score pf, preset the charging station load threshold yz, compare and analyze the charging station load threshold yz and the charging station load score pf, and evaluate the load capacity of the charging station in the future period. If the charging station load score pf is greater than or equal to the charging station load threshold yz, trigger the alarm system and perform power dispatch optimization; Step 4: Construct the objective function f(t), set up the constraint conditions, and use the gradient descent algorithm to optimize the objective function f(t) to obtain the optimal charging power P of the charging station opt ; Step 5. Based on the optimal charging power P of the charging station opt , construct a power optimization scheduling plan, monitor the power data of the charging station in real time, evaluate the effect of the power optimization scheduling, and give feedback.

Images