A dispatching control method based on grid virtual power plant optimization

Abstract

The application provides a kind of scheduling control method based on grid virtual power plant optimization, it is related to power system scheduling technical field, this method is through distributed monitoring network and collects basic data, after being preprocessed, load is predicted using deep learning model;According to the prediction result, formulate scheduling strategy, preferentially schedule the single distributed energy node of adaptation, or combine multiple nodes through collaborative scheduling algorithm;Real-time monitoring of power grid state, based on deviation classification starts dynamic adjustment mechanism;If there is still a load gap, according to the gap characteristics, access standby energy system;Finally, evaluate the scheduling effect and optimize the strategy. Through the collaborative design of accurate prediction, intelligent scheduling, dynamic adjustment and multi-objective optimization, a closed-loop system of prediction, scheduling, adjustment and optimization is formed, which comprehensively improves energy utilization efficiency, power supply reliability and scheme scientificity, adapts to high proportion of distributed energy access scenarios, and realizes efficient, stable, economic and low-carbon virtual power plant scheduling.

Description

Technical Field

[0001] This invention relates to the field of power system dispatching technology, and in particular to a dispatching control method based on power grid virtual power plant optimization. Background Technology

[0002] With the large-scale integration of distributed energy sources (such as solar and wind power), the randomness and volatility of the power grid have increased significantly, making traditional centralized dispatching methods difficult to adapt to the collaborative management needs of a high proportion of distributed energy sources. Virtual power plants, as an important technological means of integrating distributed energy sources, energy storage devices, and controllable loads, are fundamentally about achieving efficient utilization of distributed resources through optimized dispatching.

[0003] In existing technologies, virtual power plant dispatching has the following shortcomings:

[0004] 1. The collaborative scheduling algorithm does not fully consider the balance between economic efficiency and low-carbon goals, resulting in high power generation costs and carbon emissions;

[0005] 2. The dynamic adjustment mechanism has a lag in response, making it difficult to quickly restore grid stability when actual operating parameters deviate from predictions;

[0006] 3. The lack of quantitative basis for backup energy access strategies leads to low efficiency in filling load gaps and affects power supply reliability.

[0007] Therefore, there is an urgent need for a virtual power plant optimization scheduling and control method that takes into account prediction accuracy, economy, low carbon emissions, and stability in order to solve the above-mentioned technical problems. Summary of the Invention

[0008] The purpose of this invention is to provide a scheduling and control method based on the optimization of virtual power plants in the power grid, so as to improve the scheduling efficiency of virtual power plants, reduce operating costs, reduce carbon emissions, and ensure the stable operation of the power grid.

[0009] To address the aforementioned technical problems, this invention provides a dispatch control method based on power grid virtual power plant optimization, comprising the following steps:

[0010] Step 1: Collect basic data within the coverage area of ​​the virtual power plant. The basic data includes real-time load information of the main power grid, operating status parameters of each distributed energy node, charging and discharging data of energy storage devices, and external environmental influencing factors. The data is collected through a distributed monitoring network and uploaded to the control center.

[0011] Step 2: Preprocess the collected basic data and extract the feature parameters of the preprocessed basic data; construct a power grid load forecasting model based on the feature parameters to predict the load demand curve and load peak characteristics within a preset time period in the future;

[0012] Step 3: Develop a distributed energy dispatch strategy based on the load forecast results, evaluate the output capacity and response characteristics of each distributed energy node, and prioritize the dispatch of the node to the main grid if the output capacity of a single node matches the forecast load demand; if a single node cannot meet the demand, combine multiple nodes through a collaborative dispatch algorithm so that the overall output of the combined node covers the forecast load curve.

[0013] Step 4: Monitor the grid operation status in real time. When the actual operating parameters deviate from the predicted parameters, activate the dynamic adjustment mechanism and adjust the output of distributed energy nodes and the combination mode of distributed energy nodes according to the magnitude of the deviation, or call energy storage devices for power compensation.

[0014] Step 5: If a load gap still exists after dynamic adjustment, connect to the backup energy system, select the appropriate type of backup energy equipment according to the characteristics of the load gap and connect it to the grid until the grid operating parameters return to the preset range.

[0015] Preferably, the distributed energy node includes at least one of a solar power generation device, a wind power generation device, a biomass power generation device, and a small gas turbine; the external environmental influencing factors include at least one of light intensity, wind speed, temperature, humidity, and precipitation.

[0016] Preferably, the data preprocessing includes data cleaning, data standardization, and data dimensionality reduction; the characteristic parameters include load change rate, load fluctuation frequency, energy output efficiency, and energy response delay time.

[0017] Preferably, the power grid load forecasting model adopts a deep learning-based forecasting model, including an LSTM neural network model, a GRU neural network model, or a Transformer model; the input of the load forecasting model is the feature parameters extracted in step two, and the output is the load demand curve and load peak characteristics within a preset future time period;

[0018] The LSTM neural network model updates the cell state through the synergistic effect of the forget gate, input gate, and output gate. Specifically, the forget gate determines the information to be discarded from the cell state, the input gate determines the new information to be stored in the cell state, and the output gate determines the output value based on the cell state.

[0019] Preferably, in step three, the cooperative scheduling algorithm includes a scheduling algorithm based on particle swarm optimization, a scheduling algorithm based on genetic algorithm, or a scheduling algorithm based on reinforcement learning.

[0020] When the particle swarm optimization algorithm is used, its optimization objective is to minimize the comprehensive cost function, which is composed of a weighted sum of total power generation cost, total carbon emissions, and load deficit penalty term, with the sum of each weight coefficient being 1.

[0021] Preferably, in step three, when formulating a distributed energy dispatch strategy, the power generation cost, carbon emission factor, and maintenance cost of each distributed energy node should also be considered; the output capacity assessment includes the calculation of maximum output, minimum output, and economic output range, wherein the economic output range is the output range when the power generation cost is the lowest.

[0022] Preferably, in step four, the deviation between the actual operating parameters and the predicted parameters includes voltage deviation, frequency deviation, active power deviation, and reactive power deviation.

[0023] The voltage deviation is the percentage of the difference between the actual voltage value and the predicted voltage value relative to the rated voltage value;

[0024] The frequency deviation is the difference between the actual frequency value and the rated frequency value.

[0025] Preferably, in step four, the dynamic adjustment mechanism includes:

[0026] When the deviation is within the first preset threshold range, only the output of the distributed energy nodes is adjusted;

[0027] When the deviation is within the second preset threshold range, the combination method of distributed energy nodes is adjusted;

[0028] When the deviation exceeds the second preset threshold, the energy storage device is invoked for power compensation.

[0029] Preferably, in step five, the backup energy system includes a diesel generator, a natural gas generator, and an energy storage power station; the characteristics of the load gap include the gap size, duration, and time of occurrence.

[0030] Load gap characteristic quantification and judgment: Calculate the load gap size, record the duration and occurrence time. When the load gap size is greater than zero and the duration exceeds the set threshold, it is judged as a valid gap.

[0031] Backup energy equipment selection: Select equipment based on load gap threshold groups and time threshold groups, including equipment selection for short-term small gap, medium-term medium gap, and long-term large gap scenarios, as well as energy storage power stations and natural gas generators in mixed gap scenarios.

[0032] Preferably, the present invention further includes step six: evaluating the scheduling and control effect, the evaluation indicators including power supply reliability, energy utilization efficiency, operating cost and total carbon emissions;

[0033] The energy utilization efficiency is the ratio of the electrical energy effectively utilized to the total power generation.

[0034] The scheduling control strategy is optimized based on the evaluation results. The optimization process includes adjusting the parameters and weight coefficients of the cooperative scheduling algorithm.

[0035] Compared with related technologies, the scheduling and control method based on power grid virtual power plant optimization provided by this invention has the following beneficial effects:

[0036] 1. By combining distributed data acquisition with deep learning models to predict load, priority scheduling of suitable single distributed energy nodes or combination of multiple nodes using collaborative scheduling algorithms can be implemented to improve energy utilization efficiency, ensure stable power supply, reduce operating costs and carbon emissions, and form a closed-loop control system that is suitable for scenarios with a high proportion of distributed energy access.

[0037] 2. The dynamic adjustment mechanism responds in stages according to deviations: for small deviations, it adjusts the output power; for medium deviations, it adjusts the combination; and for large deviations, it calls for energy storage compensation. In conjunction with the backup energy system, it selects the appropriate type according to the gap characteristics, quickly restores grid stability, avoids excessive or insufficient adjustments, and improves power supply reliability.

[0038] 3. Optimize dispatching using a comprehensive cost function to balance power generation costs, carbon emissions, and deficit penalties. Quantify the effects through evaluation indicators and continuously optimize strategies to achieve synergy between economic efficiency and low-carbon environmental protection, thereby enhancing the scientific nature and adaptability of dispatching schemes.

[0039] In summary, through the collaborative design of accurate prediction, intelligent scheduling, dynamic adjustment, and multi-objective optimization, and by optimizing the comprehensive cost function to balance economic and low-carbon goals, a closed-loop system of prediction, scheduling, adjustment, and optimization is formed. This comprehensively improves energy utilization efficiency, power supply reliability, and the scientific nature of the solution, adapts to scenarios with a high proportion of distributed energy access, and achieves efficient, stable, economical, and low-carbon virtual power plant scheduling. Attached Figure Description

[0040] Figure 1 The flowchart of a scheduling and control method based on virtual power plant optimization in the power grid provided by the present invention is shown. Detailed Implementation

[0041] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0042] The terminology used in this disclosure is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. The singular forms “group,” “class,” and “the” as used in this disclosure and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0043] It should be understood that although the terms first, second, third, etc., may be used in this disclosure to describe various information, such information should not be limited to these terms. These terms are used only to distinguish information of the same type from one another. For example, without departing from the scope of this disclosure, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0044] Example 1

[0045] Please refer to the following: Figure 1 A dispatch control method based on grid virtual power plant optimization includes the following steps:

[0046] Step 1: Collect basic data within the coverage area of ​​the virtual power plant. The basic data includes real-time load information of the main power grid, operating status parameters of each distributed energy node, charging and discharging data of energy storage devices, and external environmental influencing factors. The data is collected through a distributed monitoring network and uploaded to the control center.

[0047] Step 2: Preprocess the collected basic data and extract the feature parameters of the preprocessed basic data; construct a power grid load forecasting model based on the feature parameters to predict the load demand curve and load peak characteristics within a preset time period in the future;

[0048] Step 3: Develop a distributed energy dispatch strategy based on the load forecast results, evaluate the output capacity and response characteristics of each distributed energy node, and prioritize the dispatch of the node to the main grid if the output capacity of a single node matches the forecast load demand; if a single node cannot meet the demand, combine multiple nodes through a collaborative dispatch algorithm so that the overall output of the combined node covers the forecast load curve.

[0049] Step 4: Monitor the grid operation status in real time. When the actual operating parameters deviate from the predicted parameters, activate the dynamic adjustment mechanism and adjust the output of distributed energy nodes and the combination mode of distributed energy nodes according to the magnitude of the deviation, or call energy storage devices for power compensation.

[0050] Step 5: If a load gap still exists after dynamic adjustment, connect to the backup energy system, select the appropriate type of backup energy equipment according to the characteristics of the load gap and connect it to the grid until the grid operating parameters return to the preset range.

[0051] This method combines distributed data acquisition with deep learning prediction models to accurately capture load demand characteristics, prioritize scheduling suitable single distributed energy nodes, or combine multiple nodes through collaborative scheduling algorithms to meet load demands. With the help of dynamic adjustment mechanisms and backup energy systems, it realizes flexible support of the virtual power plant for the main power grid. This not only improves the utilization efficiency of distributed energy and ensures power supply stability, but also reduces operating costs and carbon emissions through multi-dimensional optimization. It forms a closed-loop control system of prediction, scheduling, adjustment, and emergency response, effectively adapting to the grid scheduling needs in scenarios with a high proportion of distributed energy access.

[0052] In this application, based on step one, the distributed energy nodes include at least one of solar power generation devices, wind power generation devices, biomass power generation devices, and small gas turbines; external environmental influencing factors include at least one of light intensity, wind speed, temperature, humidity, and precipitation; the distributed monitoring network adopts a hierarchical acquisition architecture to achieve multi-level data aggregation and transmission; this part ensures comprehensive coverage of various clean and traditional distributed energy sources by clarifying the specific types of distributed energy nodes, thus improving the versatility of the method; the listed external environmental influencing factors such as light intensity and wind speed provide key input basis for subsequent load forecasting and energy dispatching, ensuring the accuracy of forecasting and dispatching; and the hierarchical acquisition architecture of the distributed monitoring network realizes efficient multi-level data aggregation and transmission, ensuring the timeliness and completeness of basic data acquisition, laying a solid data foundation for the effective operation of the entire dispatching and control method.

[0053] In this application, based on step two, data preprocessing includes data cleaning, data standardization, and data dimensionality reduction.

[0054] In this application, based on step two, the power grid load forecasting model adopts a deep learning-based forecasting model, including an LSTM neural network model, a GRU neural network model, or a Transformer model; the feature parameters of the preprocessed basic data are used as the input of the load forecasting model, and the output is the load demand curve and load peak characteristics within a future preset time period.

[0055] When using an LSTM neural network model as a power grid load forecasting model, the LSTM neural network model achieves cell state updates through the synergistic effect of the forget gate, input gate, and output gate. The specific process and calculation formula are as follows:

[0056] The forget gate is used to determine the information discarded from a cell state, and the calculation formula is as follows: ,in To preset the output coefficients of the forget gate, It is the sigmoid activation function. Here is the forget gate weight matrix. The hidden state of the previous time step. Input feature parameters for the current time step. Forget gate bias term;

[0057] The input gate is used to determine new information stored in the cell state, including:

[0058] Input gate control coefficient calculate: ,in, The input gate weight matrix, For input gate bias terms;

[0059] Candidate cell status Calculation, the calculation formula is: ,in The hyperbolic tangent activation function is used. The candidate state weight matrix is... Candidate state bias terms;

[0060] Update cell state based on the output of the forget gate and input gate. Calculation logic: In the formula, This represents the cell state at the previous time step. Represents element-wise multiplication;

[0061] The output value is determined based on the cell state, including:

[0062] Output gate control coefficient The calculation formula is as follows: ,in, This is the output gate weight matrix. This is the output gate bias term;

[0063] Hidden state The output formula is: Hidden state The output of the LSTM model is used for subsequent prediction of load demand curves and peak characteristics;

[0064] This section employs deep learning models such as LSTM, GRU, and Transformer for load forecasting, making full use of preprocessed feature parameters to improve the accuracy of predicting future load demand curves and peak characteristics. Among them, the LSTM neural network dynamically updates the cell state through the synergistic effect of the forget gate, input gate, and output gate, which can effectively capture the long-term dependencies of load sequences and solve the time series memory problem that traditional models cannot handle. This provides accurate load forecasting basis for the subsequent formulation of distributed energy dispatching strategies, ensuring the scientificity and effectiveness of the dispatching scheme.

[0065] In this application, based on step three, the cooperative scheduling algorithm includes one of the following: a particle swarm optimization-based scheduling algorithm, a genetic algorithm-based scheduling algorithm, or a reinforcement learning-based scheduling algorithm:

[0066] Obtain the unit power generation cost of any distributed energy node Planned power generation Identify the total number of distributed energy resources, n; calculate the total power generation cost by combining the unit power generation cost, planned power generation, and the total number of distributed energy resources. The formula is Where i represents the number of the distributed energy node, , Let $i$ represent the unit power generation cost and the planned power generation of the $i$-th distributed energy node, respectively.

[0067] Set the carbon emission factor for any distributed energy node Total carbon emissions are calculated using carbon emission factors and planned power generation, using the following formula: ;

[0068] The load deficit penalty is calculated by taking the difference between the actual load and actual output of the distributed energy node and the duration thereof. The formula is as follows: Where k is the preset penalty coefficient, This represents the duration of the shortfall. This represents the actual load at time t. This represents the planned output at time t;

[0069] Total generation cost, total carbon emissions, and load deficit penalty are recorded as factors affecting the generation of distributed energy nodes;

[0070] When using the particle swarm optimization algorithm, the optimization objective is to construct a function that minimizes the overall cost based on the factors affecting power generation. The function expression is as follows: ;

[0071] in , These are the weighting coefficients for total power generation cost, total carbon emissions, and load deficit penalty, respectively.

[0072] This section clarifies the selection of collaborative scheduling algorithms such as particle swarm optimization and genetic algorithms, and constructs a comprehensive cost function that includes total power generation cost, carbon emissions, and load deficit penalties by combining key parameters such as unit power generation cost and carbon emission factor. By balancing multi-objective optimization with weight coefficients, it achieves both the economic efficiency of distributed energy combined scheduling and takes into account low-carbon environmental protection and power supply reliability. It provides a scientific and quantitative decision-making basis for multi-node collaborative scheduling when a single node cannot meet load demand, and improves the rationality and feasibility of scheduling strategies.

[0073] In this application, based on step three, when formulating the distributed energy dispatch strategy, the analysis is also used to assess the generation cost, carbon emission factor, and maintenance cost of each distributed energy node; the output capacity assessment includes the calculation of maximum output, minimum output, and economic output range, wherein:

[0074] Maximum output is the maximum power generation of a distributed energy node under its rated parameters, and the calculation formula is: ;in This is the maximum output coefficient (the specific value is determined by the equipment performance). This refers to the node's rated capacity.

[0075] Minimum output is the minimum power generation of a distributed energy node during stable operation, calculated using the following formula: ,in This is the minimum output coefficient;

[0076] The economic output range is the range of power output at which the power generation cost is lowest, and the boundary values ​​of the economic output range satisfy the condition that the marginal cost is zero.

[0077] ,in Let be the generation cost function of the i-th node. This represents the output value of that node;

[0078] The power generation cost function adopts a quadratic function model. ,in , , The cost coefficients of the i-th node are respectively ( >0, ensuring the cost function is a convex function), obtained by fitting the historical power generation cost and output data of this node;

[0079] The lower limit of the economic output range is The upper limit is ,in This is the theoretical output value when the marginal cost is zero.

[0080] In this application, based on step four, the deviations between the actual operating parameters and the predicted parameters include voltage deviation, frequency deviation, active power deviation, and reactive power deviation, wherein:

[0081] The formula for calculating voltage deviation is: ;in This is the actual voltage value. To predict the voltage value, This is the rated voltage value;

[0082] The formula for calculating frequency deviation is: ;in This is the actual frequency value. This is the rated frequency value;

[0083] The formula for calculating active power deviation is: ;in This represents the actual active power. To predict active power;

[0084] The formula for calculating reactive power deviation is: ;in This represents the actual reactive power. To predict reactive power;

[0085] This section analyzes key indicators such as power generation costs and carbon emission factors of distributed energy nodes, and combines quantitative calculations of maximum output, minimum output, and economic output ranges to provide accurate references for node output capacity in scheduling strategy formulation. This ensures that scheduling schemes are both in line with equipment operating characteristics and economical. At the same time, it clarifies the calculation methods for various deviations such as voltage and frequency, providing clear quantitative standards for real-time monitoring and dynamic adjustment of grid operation status. This ensures the scientific nature of the entire process from node assessment to deviation response, and improves the accuracy and reliability of scheduling control.

[0086] In this application, the dynamic adjustment mechanism based on step four includes:

[0087] When the deviation is within the first preset threshold range, only the output of the distributed energy nodes is adjusted;

[0088] When the deviation is within the second preset threshold range, the combination method of distributed energy nodes is adjusted;

[0089] When the deviation exceeds the second preset threshold, the energy storage device is invoked for power compensation, wherein the power compensation amount of the energy storage device is calculated using a PID control algorithm.

[0090] ;in, , , These are the preset proportional, integral, and derivative coefficients, respectively;

[0091] This section establishes a dynamic adjustment mechanism with graded response by setting different adjustment strategies corresponding to different deviation threshold ranges. When the deviation is small, only the node output is adjusted; when the deviation is moderate, the node combination is adjusted; and when the deviation is large, energy storage devices are invoked and the compensation amount is accurately calculated through PID control algorithms. This ensures both the stability and flexibility of the power grid operation and allows for economical and efficient adjustment measures based on the degree of deviation, avoiding over-adjustment or under-adjustment. It provides a reliable guarantee for the rapid return of power grid parameters to the preset range.

[0092] In this application, based on step five, the backup energy system includes diesel generators, natural gas generators, and energy storage power stations; the characteristics of the load gap include the gap size, duration, and time of occurrence.

[0093] The specific process of connecting to a backup energy system includes the following sub-steps:

[0094] Sub-step 5.1, Quantification and determination of load gap characteristics:

[0095] Based on the dynamically adjusted power grid operation data from step four, the load gap is calculated using the following formula: ,in For actual load demand, For the total output of distributed energy nodes, To compensate for the power of energy storage devices;

[0096] Simultaneously, starting from the moment of dynamic adjustment failure, the time of the gap duration r and the time period of occurrence are recorded (such as peak electricity consumption period 8:00-22:00 or off-peak period 22:00-8:00 the next day); when the load gap size is greater than zero and the gap duration is greater than the set gap duration threshold, it is determined that there is an effective gap that needs to be connected to the backup energy system.

[0097] Sub-step 5.2, Selection of backup energy equipment:

[0098] Equipment is selected based on the matching degree between load gap characteristics and the technical parameters of the backup energy system, according to the following specific rules:

[0099] Set a load gap threshold group, including α and β, satisfying 0 < α < β < 1; set a gap time threshold group, including T1 and T2, satisfying 0 < T1 < T2;

[0100] When the load gap is satisfied ≤α⋅ If r≤T1, which is denoted as a short-term small gap scenario, then an energy storage power station is selected as the backup energy source for access.

[0101] When the load gap satisfies α⋅ < ≤β⋅ And if T1 < r ≤ T2, which is denoted as the mid-term gap scenario, then natural gas generators are selected as backup energy sources.

[0102] When the load gap is satisfied >β⋅ If r > T2, and this is denoted as a long-term large energy gap scenario, then a diesel generator is selected as the backup energy source.

[0103] When the load gap characteristics simultaneously meet multiple interval judgment conditions (such as...) When this occurs, the energy storage power station and the natural gas generator will trigger a coordinated response strategy.

[0104] When generating a coordinated response strategy, the energy storage power station, after identifying the load gap, prioritizes smoothing out instantaneous power fluctuations based on a preset control strategy; the natural gas generator starts with a set delay and gradually increases its output to the target value.

[0105] The total power deficit in the hybrid scenario of energy storage power stations and natural gas generators is denoted as... Perform the following power allocation:

[0106] Initial compensation power of energy storage power station: ;in, The short-term power ratio of energy storage (e.g., a value of 0.3 ≤ 0.3) ≤0.5), This refers to the rated power of the energy storage power station.

[0107] Target compensation power for natural gas generators: ,in The energy storage power attenuation coefficient (10% ≤ reserved due to the short-term discharge characteristics of energy storage). ≤20% attenuation redundancy);

[0108] Then, the deviation of key power grid parameters is dynamically maintained within a safe threshold, and adaptive adjustment is achieved through real-time data acquisition and feedback:

[0109] Frequency Deviation Control: Definition ,in For the real-time frequency of the power grid, The rated frequency;

[0110] Voltage Deviation Control: Definition ,in For the real-time voltage of the node, Rated voltage;

[0111] Based on real-time data of grid frequency, voltage, and energy storage SOC (state of charge), the power output distribution between the energy storage power station and the natural gas generator is dynamically adjusted to adapt to complex fluctuation scenarios.

[0112] This section clarifies the equipment types and quantitative judgment criteria of the backup energy system and the characteristics of the load gap. It classifies the gap scenarios and matches them with corresponding backup equipment by combining threshold groups. For mixed gap scenarios, it designs a coordinated response and power allocation strategy for energy storage and natural gas generators. At the same time, it dynamically adjusts the output by monitoring grid parameters in real time. This not only achieves accurate filling of the load gap, but also ensures the stability of grid operation. It takes into account the response efficiency and economy under different gap scenarios, and provides a scientific and comprehensive solution for the emergency dispatch of virtual power plants.

[0113] This application also includes step six: evaluating the dispatch control effect, with evaluation indicators including power supply reliability, energy utilization efficiency, operating costs, and total carbon emissions;

[0114] The formula for calculating energy efficiency is: In the formula, For electrical energy to be used effectively, To supply electricity to the total generated electricity;

[0115] Power supply reliability is expressed as power supply availability: In the formula, Available time, This refers to the total statistical period;

[0116] The scheduling control strategy is optimized based on the evaluation results. The optimization process includes adjusting the parameters and weight coefficients of the cooperative scheduling algorithm.

[0117] This section achieves a scientific measurement of the scheduling and control effect by setting evaluation indicators such as power supply reliability and energy utilization efficiency and clarifying their quantitative calculation methods. At the same time, it adjusts the parameters and weight coefficients of the collaborative scheduling algorithm based on the evaluation results, forming a closed-loop mechanism of scheduling, evaluation, and optimization. This not only verifies the actual effect of the scheduling strategy, but also continuously improves the strategy to adapt to different operating scenarios, further enhancing the economy, reliability, and low carbon emissions of the virtual power plant operation.

[0118] All formulas involved in this application are based on dimensionless numerical calculations. Dimensionlessness can be achieved through conventional means such as standardization, which will not be elaborated here. The formulas are obtained by collecting a large amount of data and simulating it with software to approximate real-world scenarios. The preset parameters are set by those skilled in the art according to the actual situation.

[0119] The above embodiments can be implemented by software, hardware, firmware, or any combination thereof. When implemented in software, it can be partially or wholly embodied as a computer program product, and when the computer instructions or programs contained therein are loaded and executed on a computer, they can generate the processes or functions of the embodiments of this application; the computer can be a general-purpose computer, a special-purpose computer, a network device, etc., and the instructions can be stored in a computer-readable medium or transmitted between media (the transmission methods include wired, wireless, etc.).

[0120] It should be understood that the sequence numbers of the processes in the various embodiments of this application do not represent the execution order. The order is determined by the function and internal logic and does not constitute a limitation on the implementation process.

[0121] Those skilled in the art will recognize that the units and algorithm steps of each example in the embodiments can be implemented by electronic hardware or a combination of hardware and software. The specific implementation method depends on the application scenario and design constraints, and its implementation should not exceed the scope of this application.

[0122] The systems, apparatuses and methods disclosed in this application are divided into units only for logical functions, and may be implemented in other ways (such as integration or splitting); mutual coupling or communication connections may be implemented through interfaces in electrical, mechanical or other forms.

[0123] Each functional unit may be integrated into the processing unit, or exist physically separately, or two or more units may be integrated into one unit. If the function is implemented in software and sold / used as an independent product, it may be stored in a computer-readable storage medium, the instructions of which may enable a computer device to execute all or part of the steps of the methods of the various embodiments of this application. The storage medium includes USB flash drives, portable hard drives, ROM, RAM, magnetic disks, optical disks, etc.

[0124] Other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of the invention are indicated by the following claims.

[0125] It should be understood that the present invention is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. A dispatch control method based on power grid virtual power plant optimization, characterized in that, Includes the following steps: Step 1: Collect basic data within the coverage area of ​​the virtual power plant. The basic data includes real-time load information of the main power grid, operating status parameters of each distributed energy node, charging and discharging data of energy storage devices, and external environmental influencing factors. The data is collected through a distributed monitoring network and uploaded to the control center. Step 2: Preprocess the collected basic data and extract the feature parameters of the preprocessed basic data; A power grid load forecasting model is constructed based on characteristic parameters to predict the load demand curve and load peak characteristics within a preset time period in the future. Step 3: Formulate a distributed energy dispatch strategy based on the load forecast results, evaluate the output capacity and response characteristics of each distributed energy node. If the output capacity of a single node matches the forecast load demand, then prioritize dispatching that node to the main grid. If a single node cannot meet the demand, then combine multiple nodes through a cooperative dispatch algorithm so that the overall output of the combined node covers the forecast load curve. The cooperative dispatch algorithm includes a dispatch algorithm based on particle swarm optimization, a dispatch algorithm based on genetic algorithm, or a dispatch algorithm based on reinforcement learning. When using the particle swarm optimization algorithm, its optimization objective is to minimize the comprehensive cost function, which is composed of a weighted sum of total power generation cost, total carbon emissions, and load deficit penalty term, with the sum of each weight coefficient being 1. Step 4: Monitor the grid operation status in real time. When the actual operating parameters deviate from the predicted parameters, activate the dynamic adjustment mechanism and adjust the output of distributed energy nodes and the combination mode of distributed energy nodes according to the magnitude of the deviation, or call energy storage devices for power compensation. Step 5: If a load gap still exists after dynamic adjustment, connect to the backup energy system, select the appropriate type of backup energy equipment according to the characteristics of the load gap and connect it to the grid until the grid operating parameters return to the preset range; The backup energy system includes diesel generators, natural gas generators, and energy storage power stations; the characteristics of the load gap include the gap size, duration, and occurrence time; load gap characteristic quantification and determination: calculate the load gap size, record the duration and occurrence time, and when the load gap size is greater than zero and the duration exceeds a set threshold, it is determined to be a valid gap; backup energy equipment selection: select equipment according to the load gap threshold group and time threshold group, including equipment selection in short-term small gap, medium-term medium gap, and long-term large gap scenarios, as well as the coordinated response strategy of energy storage power stations and natural gas generators in mixed gap scenarios.

2. The dispatch control method based on power grid virtual power plant optimization according to claim 1, characterized in that, The distributed energy node includes at least one of solar power generation devices, wind power generation devices, biomass power generation devices, and small gas turbines; the external environmental influencing factors include at least one of light intensity, wind speed, temperature, humidity, and precipitation.

3. The dispatch control method based on power grid virtual power plant optimization according to claim 1, characterized in that, The data preprocessing includes data cleaning, data standardization, and data dimensionality reduction; the characteristic parameters include load change rate, load fluctuation frequency, energy output efficiency, and energy response delay time.

4. The dispatch control method based on power grid virtual power plant optimization according to claim 1, characterized in that, The power grid load forecasting model adopts a deep learning-based forecasting model, including an LSTM neural network model, a GRU neural network model, or a Transformer model; the input of the load forecasting model is the feature parameters extracted in step two, and the output is the load demand curve and load peak characteristics within a preset future time period. The LSTM neural network model updates the cell state through the synergistic effect of the forget gate, input gate, and output gate. Specifically, the forget gate determines the information to be discarded from the cell state, the input gate determines the new information to be stored in the cell state, and the output gate determines the output value based on the cell state.

5. The dispatch control method based on power grid virtual power plant optimization according to claim 1, characterized in that, In step three, when formulating a distributed energy dispatch strategy, it is also necessary to consider the power generation cost, carbon emission factor and maintenance cost of each distributed energy node; the output capacity assessment includes the calculation of maximum output, minimum output and economic output range, where the economic output range is the output range when the power generation cost is the lowest.

6. The dispatch control method based on power grid virtual power plant optimization according to claim 1, characterized in that, In step four, the deviations between the actual operating parameters and the predicted parameters include voltage deviation, frequency deviation, active power deviation, and reactive power deviation. The voltage deviation is the percentage of the difference between the actual voltage value and the predicted voltage value relative to the rated voltage value; The frequency deviation is the difference between the actual frequency value and the rated frequency value.

7. The dispatch control method based on power grid virtual power plant optimization according to claim 1, characterized in that, In step four, the dynamic adjustment mechanism includes: When the deviation is within the first preset threshold range, only the output of the distributed energy nodes is adjusted; When the deviation is within the second preset threshold range, the combination method of distributed energy nodes is adjusted; When the deviation exceeds the second preset threshold, the energy storage device is invoked for power compensation.

8. The dispatch control method based on power grid virtual power plant optimization according to claim 1, characterized in that, It also includes step six: evaluating the effectiveness of dispatch control, with evaluation indicators including power supply reliability, energy utilization efficiency, operating costs, and total carbon emissions; The energy utilization efficiency is the ratio of the electrical energy effectively utilized to the total power generation. The scheduling control strategy is optimized based on the evaluation results. The optimization process includes adjusting the parameters and weight coefficients of the cooperative scheduling algorithm.

Images