Virtual power plant cooperative scheduling method and system based on aggregated distributed resources

Abstract

The invention relates to the technical field of smart power grids, in particular to a virtual power plant collaborative scheduling method and system based on aggregated distributed resources, and the method comprises the steps: aggregating the distributed resources, and building a model based on physical characteristics; collecting real-time state information of each resource point through a two-way communication technology, and issuing a preliminary scheduling instruction; performing local processing and preliminary adjustment on the data of each resource point, and uploading a processing result to a cloud for global optimization; constructing an optimal scheduling model, and dynamically adjusting a scheduling strategy through an intelligent algorithm to obtain a dynamic scheduling scheme; feedback data in the scheduling execution process are obtained, parameters of the scheduling model are adjusted and optimized in real time, and iterative optimization is conducted on the dynamic scheduling scheme. According to the invention, the problems that the traditional energy system is independent and the scheduling real-time performance and accuracy are low are effectively solved, the unified scheduling and real-time optimization of the distributed resources are realized, and the energy management and optimization capability of the virtual power plant is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of smart grid technology, and in particular to a virtual power plant collaborative scheduling method and system based on aggregated distributed resources. Background Art

[0002] The current energy system mainly includes gas turbine equipment, waste heat boiler equipment, electric refrigeration units, flue gas hot water lithium bromide units, gas boilers, photovoltaic systems, energy storage systems and other equipment. Each system is independent and difficult to coordinate and control resources. The production plan scheduling and regional scheduling are relatively simple and mainly rely on manual scheduling. This scheduling method has low real-time and accuracy and cannot achieve overall linkage adjustment and control. Without accurate data analysis and algorithm prediction of the efficiency and accuracy of resource allocation, the energy management and optimization capabilities of virtual power plants are not high. Summary of the invention

[0003] The present invention provides a virtual power plant collaborative scheduling method and system based on aggregated distributed resources, thereby effectively solving the problems pointed out in the background technology.

[0004] In order to achieve the above object, the technical solution adopted by the present invention is:

[0005] A virtual power plant collaborative dispatching method based on aggregated distributed resources, comprising:

[0006] Aggregate distributed resources, build models based on physical characteristics, and quantify the adjustable capacity and adjustment rate of each resource;

[0007] Collect real-time status information of each resource point through two-way communication technology and issue preliminary dispatch instructions;

[0008] Based on edge computing technology, the data of each resource point is locally processed and preliminarily adjusted, and the processing results are uploaded to the cloud for global optimization;

[0009] Based on the power system operating parameters, an optimization dispatch model is constructed, and the dispatch strategy is dynamically adjusted through intelligent algorithms to obtain a dynamic dispatch solution;

[0010] Acquire feedback data during the scheduling execution process, adjust the parameters of the optimization scheduling model in real time, and iteratively optimize the dynamic scheduling solution.

[0011] Furthermore, the photovoltaic resource model includes:

[0012] PV power generation model, I ph =G / G ref ×(I phref +k×(T cell -T ref)),

[0013] Battery model, V cell =V oc +m×(T cell -T ref )-I cell ×R cell ,

[0014] Inverter model, P ac =η×P dc -P loss -P mppt ,

[0015] Among them, I ph is the short-circuit current of the photovoltaic cell, G is the solar radiation intensity, G ref The standard radiation intensity (1000W / m 2 ), I phref is the standard short-circuit current (A), T cell is the photovoltaic cell temperature, T ref is the standard temperature (25°C), k is the photovoltaic cell temperature coefficient, V cell Photovoltaic cell voltage, V oc is the open circuit voltage, m is the temperature coefficient, I cell is the battery current, R cell is the internal resistance of the battery, P ac is the inverter output power, P dc is the DC power of the photovoltaic array, η is the conversion efficiency of the inverter, P loss is the power loss of the inverter, P mppt is the power loss of the maximum power point tracker (MPPT).

[0016] Furthermore, the energy storage resource model includes:

[0017] Battery Model,

[0018]

[0019] U cell (S, T) = U0(T) - IR i (S, T)+K4T ln(S)+K5T ln(1-S),

[0020] Among them, S is the battery SOC, S0 is the initial SOC value, C cell is the capacity of a single battery (in Ah), n is the proportional constant taking into account the number of batteries in parallel and the battery charge and discharge efficiency, I is the charge and discharge current (in kA), t is the integral variable (in s), S age.loss and C age.lossare the SOC loss and capacity decay caused by battery aging, S T.loss and C T.loss are the SOC loss and capacity attenuation caused by temperature, m is the number of charge and discharge cycles, U cell (S, T) is the battery cell voltage (in V), U0 is the initial battery voltage (in V), R is the battery internal resistance, T is the temperature, K 1(m) and K(T) to K0(T) are correction coefficients considering the effects of cycle number and temperature, respectively.

[0021] Furthermore, the charging pile model includes:

[0022]

[0023] Where k is the current cycle, n is the number of charging piles involved in voltage regulation, if there is a node connected to multiple charging piles, it will be split into multiple nodes so that each node is connected to at most one charging pile, therefore, n is also the number of nodes involved in voltage regulation, k is the number of predicted cycles, is the predicted value of the voltage at the node where the charging pile j is located at the initial time of the future k+h cycle, V0 is the voltage target value, U all is the average capacity utilization rate of all charging piles in the area, U j is the capacity utilization rate of charging pile j, P j , Q j , They are the active power, reactive power and rated capacity of the charging pile respectively.

[0024] Furthermore, after collecting the real-time status information of the resource points and issuing the preliminary dispatch instructions, and before the local processing and preliminary adjustment of the resource point data, the response capability is accurately quantified, including:

[0025] Obtain the capacity declared by the user and verify it according to the capacity constraints;

[0026] Obtain the actual corresponding capacity of adjacent time periods and determine whether the adjustment rate constraints are met;

[0027] Based on the actual corresponding capacity, the corresponding time delay and the corresponding error are corrected, and the effective corresponding capacity is calculated.

[0028] Furthermore, the real-time status information of each resource point is collected through two-way communication technology, and preliminary dispatch instructions are issued, including:

[0029] Using two-way communication technology, key status information of each distributed resource point is collected in real time;

[0030] Pre-process the collected status information and monitor the operating status of each resource point in real time;

[0031] Based on the pre-processed status information and historical data, the prediction algorithm is used to predict the energy demand in the future, and a preliminary scheduling strategy is formulated according to the prediction results and the real-time status of each resource point;

[0032] The preliminary scheduling strategy is converted into specific scheduling instructions and sent to each distributed resource point through two-way communication technology.

[0033] Furthermore, the intelligent algorithm includes:

[0034] Construct an optimization model for the operating cost of the virtual power plant system, where the operating cost is the minimum sum of the electricity purchase cost and the gas cost;

[0035] Based on the optimization model, the electricity purchase cost and gas cost are calculated as follows: and Where i is the number of scheduling period periods, is the power purchased from the grid during period t, represents the electricity purchase price in time period t, i is the number of time periods in the dispatch cycle, is the power generation of the gas internal combustion engine during period t, c fc represents the natural gas price, F fuel is the calorific value of natural gas, η GE is the gas power generation efficiency, Δt is the time period length;

[0036] The operating cost of the virtual power plant system is calculated based on the electricity purchase cost and gas cost. The formula is:

[0037] Furthermore, the data of each resource point is locally processed and preliminarily adjusted based on edge computing technology, and the processing results are uploaded to the cloud for global optimization, including:

[0038] Preprocess the collected real-time operation data of distributed resources and calculate the status evaluation index based on voltage deviation, frequency deviation and power deviation;

[0039] Determine whether to trigger a quick response based on the status assessment results. When the status indicator exceeds the preset threshold, select a solution from the preset control library for local preliminary adjustment.

[0040] The processed data is compressed by extracting key feature parameters and using a time-space compression algorithm, and the compressed data is uploaded to the cloud in batches;

[0041] Perform multivariate data alignment and fusion on the uploaded data of each edge node, establish a global state evaluation model, and build an optimization objective function based on the evaluation results to solve the global optimal solution;

[0042] The optimized control strategy is sent to each edge node for state fine-tuning, and the execution result is fed back to the cloud.

[0043] A virtual power plant collaborative dispatching system based on aggregated distributed resources, the system comprising:

[0044] The physical model building module aggregates distributed resources, builds a model based on physical characteristics, and quantifies the adjustable capacity and adjustment rate of each resource;

[0045] The preliminary instruction issuing module collects the real-time status information of each resource point through two-way communication technology and issues preliminary dispatching instructions;

[0046] The edge computing optimization module processes and preliminarily adjusts the data of each resource point locally based on edge computing technology, and uploads the processing results to the cloud for global optimization;

[0047] The dispatch scheme acquisition module builds an optimized dispatch model based on the power system operating parameters, and dynamically adjusts the dispatch strategy through artificial intelligence algorithms to obtain a dynamic dispatch scheme;

[0048] The scheme iteration optimization module obtains feedback data during the scheduling execution process, adjusts the parameters of the optimization scheduling model in real time, and iteratively optimizes the dynamic scheduling scheme.

[0049] Furthermore, the edge computing optimization module includes:

[0050] An evaluation index calculation unit pre-processes the collected real-time operation data of distributed resources and calculates the state evaluation index based on voltage deviation, frequency deviation and power deviation;

[0051] The local preliminary adjustment unit determines whether to trigger a quick response based on the status evaluation results. When the status indicator exceeds the preset threshold, a solution is selected from the preset control library for local preliminary adjustment.

[0052] The data compression upload unit compresses the processed data by extracting key feature parameters and using a time-space compression algorithm, and uploads the compressed data to the cloud in batches;

[0053] The optimal solution solving unit performs multivariate data alignment and fusion on the uploaded data of each edge node, establishes a global state evaluation model, and constructs an optimization objective function based on the evaluation results to solve the global optimal solution;

[0054] The strategy fine-tuning feedback unit sends the optimized control strategy to each edge node for state fine-tuning, and feeds back the execution result to the cloud.

[0055] The technical solution of the present invention can achieve the following technical effects:

[0056] It effectively solves the problems of traditional energy systems operating independently and low scheduling real-time and accuracy, realizes unified scheduling and real-time optimization of distributed resources, and significantly improves the energy management and optimization capabilities of virtual power plants. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0058] Figure 1 A flowchart of a collaborative scheduling method for a virtual power plant based on aggregated distributed resources;

[0059] Figure 2 Schematic diagram of the process for accurate quantification of response capacity;

[0060] Figure 3 A flow chart for collecting real-time status information of each resource point and issuing preliminary dispatch instructions;

[0061] Figure 4 Schematic diagram of the process for local processing and preliminary adjustments and uploading to the cloud for global optimization. DETAILED DESCRIPTION

[0062] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0063] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items.

[0064] Embodiment 1

[0065] like Figure 1 As shown, the present invention provides a virtual power plant collaborative scheduling method based on aggregated distributed resources, the method comprising:

[0066] S1: Aggregate distributed resources, build a model based on physical characteristics, and quantify the adjustable capacity and adjustment rate of each resource;

[0067] Specifically, distributed energy resources such as photovoltaics, wind power, energy storage, and electric vehicle charging piles are aggregated to form a virtual overall resource pool. For each type of resource, a physical property model is established. For example, the model of the photovoltaic system takes into account the impact of solar radiation intensity and temperature on power generation, and the model of the energy storage system includes charging and discharging characteristics, SOC (state) changes, aging effects, etc. The regulation capability of each resource is analyzed through the precise quantification technology algorithm of response capability. For example, the adjustable capacity of the photovoltaic system is determined according to the real-time solar radiation intensity, and the regulation rate of the energy storage system is based on the current charge and discharge state and remaining capacity of the battery.

[0068] S2: Collect real-time status information of each resource point through two-way communication technology and issue preliminary dispatch instructions;

[0069] Specifically, the operating status of distributed resources is obtained in real time, such as the output power of photovoltaic power generation, the charging and discharging status of energy storage batteries, the usage of electric vehicle charging piles, etc., so as to grasp the current status of each resource; at the same time, preliminary scheduling instructions are issued through two-way communication, and resources can be preliminarily adjusted according to the collected real-time status to achieve rapid response and local optimization, ensuring that the system is flexible and real-time when facing load fluctuations.

[0070] S3: Based on edge computing technology, the data of each resource point is processed locally and preliminarily adjusted, and the processing results are uploaded to the cloud for global optimization;

[0071] Specifically, through edge computing technology, efficient management and real-time response of various distributed resources can be achieved. First, real-time data collection is carried out near the resource point, the acquired information is quickly processed, and data filtering, denoising and outlier detection are performed to ensure data accuracy; then, according to the processed real-time data, the resources are initially adjusted, such as adjusting the charging and discharging strategy of energy storage equipment or controlling the output of photovoltaic and wind energy systems to cope with load fluctuations. This local adjustment can quickly respond to changes and enhance the stability of the power grid. By performing preliminary processing locally, the amount of data that needs to be transmitted to the cloud is reduced, and the network bandwidth requirements and latency are reduced. The processing results will be uploaded to the cloud to provide accurate information for global optimization and help build a comprehensive power system operation model. The cloud uses this data for global optimization scheduling, formulates refined scheduling plans, and improves the operating efficiency and economy of the virtual power plant, thereby ensuring the stability and flexibility of the entire system.

[0072] S4: Based on the power system operating parameters, an optimization dispatch model is constructed, and the dispatch strategy is dynamically adjusted through intelligent algorithms to obtain a dynamic dispatch solution;

[0073] Specifically, firstly, various real-time data are collected and integrated, including power demand forecasts, actual output of renewable energy, power market prices, and user demand response information; then, these data are used to build a comprehensive optimization dispatch model, which can reflect the interactive relationship between different resources and the operating characteristics of the power grid; next, the optimization dispatch model is solved by intelligent algorithms, and the dispatch strategy is adjusted dynamically. The intelligent algorithm will optimize the resource dispatch plan according to the real-time system status and market changes, ensure the priority use of renewable energy, balance the supply and demand relationship, and reduce the system operation cost; finally, the generated dynamic dispatch plan will be sent to each distributed resource point to guide the actual operation. The flexibility and adaptability of this plan enable the power system to achieve efficient and stable operation in the face of uncertainty and change.

[0074] S5: Obtain feedback data during the scheduling execution process, adjust the parameters of the optimization scheduling model in real time, and iteratively optimize the dynamic scheduling solution.

[0075] Specifically, during the dispatch execution process, feedback data from each resource point will be continuously monitored and collected. These feedback data include actual power generation, load demand, resource response, and overall operating status. These real-time data can accurately evaluate the effectiveness of the dispatch strategy and the actual performance of the resources; after collecting the feedback data, the parameters of the optimization dispatch model will be adjusted in real time. For example, if the actual output of a resource is found to be lower than expected, the reasons (such as weather changes, equipment failures, etc.) will be analyzed and the adjustment capacity parameters of the resource will be adjusted accordingly. In addition, based on factors such as load changes and market price fluctuations, the dispatch strategy parameters in the model will also be optimized accordingly; the adjusted optimization dispatch model will be used to iteratively optimize the dynamic dispatch scheme. This process ensures that the dispatch scheme can continuously adapt to real-time demand and environmental changes. Through repeated feedback and optimization, more efficient and reliable resource dispatch can be achieved under the ever-changing power market and environmental conditions, thereby improving the flexibility and stability of the overall power system.

[0076] The present invention effectively solves the problems of traditional energy systems operating independently and low scheduling real-time and accuracy, realizes unified scheduling and real-time optimization of distributed resources, and significantly improves the energy management and optimization capabilities of virtual power plants.

[0077] As a preferred embodiment of the above, the photovoltaic resource model includes:

[0078] PV power generation model, I ph =G / Gref ×(I phref +k×(T cell -T ref )),

[0079] Battery model, V cell =V oc +m×(T cell -T ref )-I cell ×R cell ,

[0080] Inverter model, P ac =η×P dc -P loss -P mppt ,

[0081] Among them, I ph is the short-circuit current of the photovoltaic cell, G is the solar radiation intensity, G ref The standard radiation intensity (1000W / m 2 ), I phref is the standard short-circuit current (A), T cell is the photovoltaic cell temperature, T ref is the standard temperature (25°C), k is the photovoltaic cell temperature coefficient, V cell Photovoltaic cell voltage, V oc is the open circuit voltage, m is the temperature coefficient, I cell is the battery current, R cell is the internal resistance of the battery, P ac is the inverter output power, P dc is the DC power of the photovoltaic array, η is the conversion efficiency of the inverter, P loss is the power loss of the inverter, P mppt is the power loss of the maximum power point tracker (MPPT).

[0082] Specifically, in the photovoltaic resource model, the meaning of each variable is crucial. First, Iph represents the short-circuit current of the photovoltaic cell, which indicates the maximum current that can be output when the photovoltaic cell is short-circuited. This value is usually proportional to the light intensity. The actual solar radiation intensity G refers to the actual light intensity irradiated on the photovoltaic cell, expressed in watts per square meter (W / m 2 ) is the unit, and G ref is the standard radiation intensity, set to 1000W / m 2 , represents the reference light intensity of the photovoltaic cell under standard test conditions, the standard short-circuit current I phref is the short-circuit current of the photovoltaic cell under the reference light intensity, in amperes (A), and the actual temperature of the photovoltaic cell T cellIt will affect its performance and is usually expressed in degrees Celsius (℃), while the standard temperature T cell It is set to 25°C for comparison under standard test conditions. The temperature coefficient k represents the effect of temperature change on the short-circuit current and is usually a negative value, indicating that the current will decrease when the temperature increases.

[0083] In the battery model, the photovoltaic cell voltage V cell Indicates the output voltage of the battery in working state, while the open circuit voltage V oc It is the voltage of the photovoltaic cell under no-load conditions, indicating the maximum value of the voltage across the battery. The temperature coefficient m indicates the effect of temperature change on the battery voltage, which is usually a negative value, meaning that the voltage will decrease when the temperature rises. The battery current I cell is the actual current passing through the battery, reflecting the output of the battery under load, while the battery internal resistance R cell It indicates the internal impedance of the battery, which affects the output power and efficiency of the battery.

[0084] In the inverter model, the output power P ac Represents the power output after the inverter converts DC power into AC power, in watts (W). dc Refers to the DC power generated by the photovoltaic cell array under specific conditions. The inverter conversion efficiency η is a value between 0 and 1, indicating the efficiency of the inverter in converting DC power to AC power, usually expressed as a percentage. The inverter power loss P loss The power loss is caused by internal resistance, heat and other factors, while the power loss P of the maximum power point tracker (MPPT) is mppt It represents the power loss caused by the control circuit when tracking the maximum power point.

[0085] As a preferred embodiment of the above, the energy storage resource model includes:

[0086] Battery Model,

[0087]

[0088] U cell (S, T) = U0(T) - IR i (S, T)+K4T ln(S)+K5T ln(1-S),

[0089] Among them, S is the battery SOC, S0 is the initial SOC value, C cell is the capacity of a single battery (in Ah), n is the proportional constant taking into account the number of batteries in parallel and the battery charge and discharge efficiency, I is the charge and discharge current (in kA), t is the integral variable (in s), S age.loss and Cage.loss are the SOC loss and capacity decay caused by battery aging, S T.loss and C T.loss are the SOC loss and capacity attenuation caused by temperature, m is the number of charge and discharge cycles, U cell (S, T) is the battery cell voltage (in V), U0 is the initial battery voltage (in V), R is the battery internal resistance, T is the temperature, K 1(m) and K(T) to K0(T) are correction coefficients considering the effects of cycle number and temperature, respectively.

[0090] Specifically, first, the battery state (SOC, State of Charge) is represented by S, and the initial state is S0. The basic equation of the battery model is: In this equation, C cell is the capacity of a single battery (in ampere-hours, Ah), and n is a constant. Considering the number of parallel batteries and the charging and discharging efficiency, the integral term ∫I(t)dt represents the accumulation of charging and discharging current over time, reflecting the change in SOC of the battery during use. T.loss and C T.loss They represent the SOC loss due to battery aging and temperature change, respectively, which further affects the actual available capacity of the battery.

[0091] The aging loss is calculated by the following formula: Here, C age.loss represents the capacity decay due to battery aging, while K 1(m) It is a correction factor that takes into account the effect of the number of charge and discharge cycles m on battery performance, indicating that the battery capacity will gradually decrease with the increase of usage cycles.

[0092] The temperature effect loss is expressed by the following formula: Here, C T.loss It is the capacity attenuation caused by temperature change. K2(T) is the correction factor considering the temperature effect. It reflects the impact of temperature on battery performance. As the temperature changes, the effective capacity and charging capacity of the battery will also change.

[0093] Finally, the battery cell voltage U cell The calculation formula of (S, T) is: cell (S, T) = U0(T) - IR i (S, T) + K4Tln(S) + K5T ln(1-S) In this equation, U0(T) is the voltage of the battery in the initial state, I is the charge and discharge current, and R i(S, T) represents the internal resistance of the battery, which may change with changes in SOC and temperature. The temperature-related correction coefficients K4 and K5 affect the performance of the battery voltage. Through these formulas, the energy storage resource model can dynamically describe the battery's SOC changes, capacity loss and voltage characteristics, providing support for the scheduling decisions of the virtual power plant.

[0094] As a preferred embodiment of the above, the charging pile model includes:

[0095]

[0096] Where k is the current cycle, n is the number of charging piles involved in voltage regulation, if there is a node connected to multiple charging piles, it will be split into multiple nodes so that each node is connected to at most one charging pile, therefore, n is also the number of nodes involved in voltage regulation, k is the number of predicted cycles, is the predicted value of the voltage at the node where the charging pile j is located at the initial time of the future k+h cycle, V0 is the voltage target value, U all is the average capacity utilization rate of all charging piles in the area, U j is the capacity utilization rate of charging pile j, P j , Q j , S j max They are the active power, reactive power and rated capacity of the charging pile respectively.

[0097] Specifically, first, the objective function in the model is:

[0098] The purpose of this formula is to minimize the sum of the squares of the voltage deviations of all charging piles involved in voltage regulation. Through this objective function, the output of the charging piles can be effectively adjusted to ensure that the voltage remains within the target range, thereby enhancing the stability of the power system. Secondly, the average capacity utilization rate of the charging piles U all is defined as: In this formula, n is the number of charging piles involved in voltage regulation (each node can be connected to at most one charging pile), U j is the capacity utilization rate of charging pile j, P j and Q j They represent the active power and reactive power of the charging pile respectively, and S j max is the rated capacity of the charging pile. By calculating the average utilization rate of all charging piles, the overall operating efficiency of the charging piles in the region can be evaluated.

[0099] As a preferred embodiment of the above, after collecting the real-time status information of the resource point and issuing the preliminary scheduling instruction, and before the resource point data is locally processed and preliminarily adjusted, the response capability is accurately quantified, including:

[0100] A10: Obtain the capacity declared by the user and perform verification according to the capacity constraint conditions;

[0101] A20: Obtain the actual corresponding capacity of adjacent time periods and determine whether the adjustment rate constraint is met;

[0102] A30: Based on the actual corresponding capacity, the corresponding time delay and the corresponding error are corrected, and the effective corresponding capacity is calculated.

[0103] Specifically, first, obtain the declared capacity of each user, which is usually the power capacity that the user hopes to provide or consume in a specific time period. This data comes from the user's demand forecast or historical power consumption data; then, check these declared capacities based on capacity constraints (such as grid load capacity, resource availability, etc.) to ensure that they are within the permitted range. This step ensures the feasibility and safety of the dispatch process and avoids grid instability caused by excessive declaration; next, obtain the actual response capacity of adjacent time periods, which refers to the power capacity that the user can actually respond to in the previous time period, and evaluate whether these actual response capacities meet the regulation rate constraints, that is, how quickly the resources can increase or decrease the output power within the specified time. This judgment is crucial to ensure the stable operation of the power system, especially when the load fluctuates greatly; finally, based on the actual response capacity obtained, correct the corresponding time delay and response error, which includes considering the delay factors that may exist in actual operation, such as equipment response time, communication delay, etc., as well as the response error caused by measurement error or inaccurate model. Through these corrections, the effective response capacity is calculated, which represents the response capability that the user can actually provide during the dispatch process.

[0104] As a preferred embodiment of the above embodiment, the real-time status information of each resource point is collected through a two-way communication technology, and a preliminary scheduling instruction is issued, including:

[0105] S21: Using two-way communication technology, collect key status information of each distributed resource point in real time;

[0106] S22: pre-processing the collected status information and monitoring the operating status of each resource point in real time;

[0107] S23: Based on the pre-processed status information and historical data, the prediction algorithm is used to predict the energy demand in the future period, and a preliminary scheduling strategy is formulated according to the prediction results and the real-time status of each resource point;

[0108] S24: Convert the preliminary scheduling strategy into specific scheduling instructions, and send them to each distributed resource point through two-way communication technology.

[0109] Specifically, firstly, the key status information of each distributed resource point is collected in real time by using two-way communication technology. This information may include power generation, energy storage status, load demand and equipment operation status. Through this process, a comprehensive and timely understanding of the operation status of each resource point can be ensured; the collected status information is preprocessed, including data cleaning, denoising and outlier detection. By preprocessing the data, the data quality can be improved to ensure the accuracy of subsequent analysis. At the same time, the operation status of each resource point is monitored in real time, potential problems are found and timely feedback is given to ensure the normal operation of resources; based on the preprocessed status information and historical data, the prediction algorithm is used to predict the energy demand in the future period. This prediction can be combined with machine learning or statistical models to improve the accuracy of the prediction. According to the prediction results and the real-time status of each resource point, a preliminary scheduling strategy is formulated to optimize the use of resources and meet future energy needs; finally, the preliminary scheduling strategy is converted into specific scheduling instructions. These instructions may include specific requirements for power generation, energy storage operation and load distribution. These instructions are sent to each distributed resource point through two-way communication technology to ensure that resources can be adjusted and responded according to the scheduling strategy, so as to achieve the balance and stability of the overall power grid.

[0110] As a preferred embodiment of the above, the intelligent algorithm includes:

[0111] Construct an optimization model for the operating cost of the virtual power plant system, where the operating cost is the minimum sum of the electricity purchase cost and the gas cost;

[0112] Based on the optimization model, the electricity purchase cost and gas cost are calculated as follows: and Where i is the number of scheduling period periods, is the power purchased from the grid during period t, represents the electricity purchase price in time period t, i is the number of time periods in the dispatch cycle, is the power generation of the gas internal combustion engine during period t, c fc represents the natural gas price, F fuel is the calorific value of natural gas, η GE is the gas power generation efficiency, Δt is the time period length;

[0113] The operating cost of the virtual power plant system is calculated based on the electricity purchase cost and gas cost. The formula is:

[0114]

[0115] Specifically, the total operating cost C is composed of the electricity purchase cost C te and gas cost C tf The formula for calculating the electricity purchase cost is: This formula is calculated by calculating the power purchase price C in each dispatch cycle. te And the power C purchased from the grid te The total electricity purchase cost is obtained by adding up the time period length Δt. This calculation ensures the economic efficiency of electricity purchase while meeting the electricity demand. On the other hand, the calculation of gas cost is expressed by the following formula: In this formula, c fc is the price of natural gas, is the power generated by the gas internal combustion engine in time period t, and η GE It represents the efficiency of gas-fired power generation. By converting the energy consumption of gas-fired power generation into costs and summing them up, the economic feasibility of gas-fired power generation can be accurately evaluated.

[0116] Ultimately, the total operating cost of a virtual power plant can be calculated using the following comprehensive formula: This formula combines the costs of purchasing electricity and gas to derive the overall operating expenses of the virtual power plant in different time periods. This comprehensive cost calculation helps optimize the scheduling strategy and improve the economy and operating efficiency of the virtual power plant, thereby minimizing operating costs while meeting electricity demand.

[0117] As a preferred embodiment of the above, the data of each resource point is locally processed and preliminarily adjusted based on edge computing technology, and the processing results are uploaded to the cloud for global optimization, including:

[0118] S31: preprocessing the collected real-time operation data of distributed resources, and calculating a state evaluation index based on voltage deviation, frequency deviation, and power deviation;

[0119] S32: judging whether to trigger a quick response based on the status evaluation result, and selecting a solution from a preset control library for local preliminary adjustment when the status indicator exceeds a preset threshold;

[0120] S33: compressing the processed data by extracting key feature parameters and using a spatiotemporal compression algorithm, and uploading the compressed data to the cloud in batches;

[0121] S34: perform multivariate data alignment and fusion on the uploaded data of each edge node, establish a global state evaluation model, and construct an optimization objective function based on the evaluation results to solve the global optimal solution;

[0122] S35: Send the optimized control strategy to each edge node for state fine-tuning, and feed back the execution results to the cloud.

[0123] Specifically, first, the collected real-time operation data of distributed resources are preprocessed. This process includes data cleaning and denoising to ensure data quality. After preprocessing, the state evaluation indicators are calculated based on key indicators such as voltage deviation, frequency deviation and power deviation. These indicators reflect the operation status of the resource point and the degree of deviation from the preset target. Then, according to the state evaluation results, it is determined whether to trigger a rapid response. When the state evaluation indicator exceeds the preset threshold, it means that the operation status of the resource point is abnormal and needs to be adjusted. In this case, an appropriate adjustment scheme is selected from the preset control library to perform local preliminary adjustment to ensure the stable operation of the resources. Then, the key feature parameters are extracted and the processed data are compressed using the time-space compression algorithm. The data is compressed, which can reduce the amount of data that needs to be uploaded and improve data transmission efficiency. The compressed data will be uploaded to the cloud in batches to ensure the timeliness and integrity of the data. In the cloud, the data uploaded by each edge node is aligned and fused with multiple data. By establishing a global state evaluation model, the information of each resource point is integrated to comprehensively evaluate the overall operation status of the system. Based on these evaluation results, the optimization objective function is constructed to solve the global optimal solution, thereby formulating the best scheduling strategy. Finally, the optimized control strategy is sent to each edge node for state fine-tuning. The edge node adjusts its own operation status according to the new control instructions and feeds back the execution results to the cloud for further analysis and optimization.

[0124] Embodiment 2

[0125] Based on the same inventive concept as the virtual power plant collaborative dispatching method based on aggregated distributed resources in the aforementioned embodiment, the present invention further provides a virtual power plant collaborative dispatching system based on aggregated distributed resources, the system comprising:

[0126] The physical model building module aggregates distributed resources, builds a model based on physical characteristics, and quantifies the adjustable capacity and adjustment rate of each resource;

[0127] The preliminary instruction issuing module collects the real-time status information of each resource point through two-way communication technology and issues preliminary dispatching instructions;

[0128] The edge computing optimization module processes and preliminarily adjusts the data of each resource point locally based on edge computing technology, and uploads the processing results to the cloud for global optimization;

[0129] The dispatch scheme acquisition module builds an optimized dispatch model based on the power system operating parameters, and dynamically adjusts the dispatch strategy through artificial intelligence algorithms to obtain a dynamic dispatch scheme;

[0130] The scheme iteration optimization module obtains feedback data during the scheduling execution process, adjusts the parameters of the optimization scheduling model in real time, and iteratively optimizes the dynamic scheduling scheme.

[0131] The above-mentioned dispatching system in the present invention can effectively implement the collaborative dispatching method of virtual power plants based on aggregated distributed resources. The technical effects that can be achieved are as described in the above-mentioned embodiments and will not be repeated here.

[0132] As a preferred embodiment of the above, the edge computing optimization module includes:

[0133] An evaluation index calculation unit pre-processes the collected real-time operation data of distributed resources and calculates the state evaluation index based on voltage deviation, frequency deviation and power deviation;

[0134] The local preliminary adjustment unit determines whether to trigger a quick response based on the status evaluation results. When the status indicator exceeds the preset threshold, a solution is selected from the preset control library for local preliminary adjustment.

[0135] The data compression upload unit compresses the processed data by extracting key feature parameters and using a time-space compression algorithm, and uploads the compressed data to the cloud in batches;

[0136] The optimal solution solving unit performs multivariate data alignment and fusion on the uploaded data of each edge node, establishes a global state evaluation model, and constructs an optimization objective function based on the evaluation results to solve the global optimal solution;

[0137] The strategy fine-tuning feedback unit sends the optimized control strategy to each edge node for state fine-tuning and feeds back the execution results to the cloud.

[0138] Similarly, the above-mentioned optimization schemes for the system can also respectively achieve the optimization effects corresponding to the method in Example 1, which will not be repeated here.

[0139] Although the present application has been described in conjunction with specific features and embodiments thereof, it is obvious that various modifications and combinations may be made thereto without departing from the spirit and scope of the present application. Accordingly, this specification and the accompanying drawings are merely exemplary illustrations of the present application as defined therein, and are deemed to have covered any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application is intended to include these modifications and variations.

Claims

1. A virtual power plant collaborative scheduling method based on aggregated distributed resources, characterized in that: include: Aggregate distributed resources, build models based on physical characteristics, and quantify the adjustable capacity and adjustment rate of each resource; Collect real-time status information of each resource point through two-way communication technology and issue preliminary dispatch instructions; Based on edge computing technology, the data of each resource point is locally processed and preliminarily adjusted, and the processing results are uploaded to the cloud for global optimization; Based on the power system operating parameters, an optimization dispatch model is constructed, and the dispatch strategy is dynamically adjusted through intelligent algorithms to obtain a dynamic dispatch solution; Acquire feedback data during the scheduling execution process, adjust the parameters of the optimization scheduling model in real time, and iteratively optimize the dynamic scheduling solution.

2. The method for collaborative scheduling of virtual power plants based on aggregated distributed resources according to claim 1 is characterized in that: The photovoltaic resource model includes: PV power generation model, I ph =G / G ref ×(I phref +k×(T cell -T ref )), Battery model, V cell =V oc +m×(T cell -T ref )-I cell ×R cell , Inverter model, P ac =η×P dc -P loss -P mppt , Among them, I ph is the short-circuit current of the photovoltaic cell, G is the solar radiation intensity, G ref The standard radiation intensity (1000W / m 2 ), I phref is the standard short-circuit current (A), T cell is the photovoltaic cell temperature, T ref is the standard temperature (25°C), k is the photovoltaic cell temperature coefficient, V cell Photovoltaic cell voltage, V oc is the open circuit voltage, m is the temperature coefficient, I cell is the battery current, R cell is the internal resistance of the battery, P ac is the inverter output power, P dc is the DC power of the photovoltaic array, η is the conversion efficiency of the inverter, P loss is the power loss of the inverter, P mppt is the power loss of the maximum power point tracker (MPPT).

3. The method for collaborative scheduling of virtual power plants based on aggregated distributed resources according to claim 1, characterized in that: The energy storage resource model includes: Battery Model, U cell (S,T)=U0(T)-IR i (S,T)+K4Tln(S)+K5Tln(1-S), Among them, S is the battery SOC, S0 is the initial SOC value, C cell is the capacity of a single battery (in Ah), n is the proportional constant taking into account the number of batteries in parallel and the battery charge and discharge efficiency, I is the charge and discharge current (in kA), t is the integral variable (in s), S age.loss and C age.loss are the SOC loss and capacity decay caused by battery aging, S T.loss and C T.loss are the SOC loss and capacity attenuation caused by temperature, m is the number of charge and discharge cycles, U cell (S, T) is the battery cell voltage (in V), U0 is the initial state battery voltage (in V), R is the battery internal resistance, T is the temperature, K 1(m) and K(T) to K0(K) are correction coefficients considering the effects of cycle number and temperature, respectively.

4. The method for collaborative scheduling of virtual power plants based on aggregated distributed resources according to claim 1, characterized in that: Charging pile models include: Where k is the current cycle, n is the number of charging piles involved in voltage regulation, if there is a node connected to multiple charging piles, it will be split into multiple nodes so that each node is connected to at most one charging pile, therefore, n is also the number of nodes involved in voltage regulation, k is the number of predicted cycles, is the predicted value of the voltage at the node where the charging pile j is located at the initial time of the future k+h cycle, V0 is the voltage target value, U all is the average capacity utilization rate of all charging piles in the area, U j is the capacity utilization rate of charging pile j, P j , Q j , They are the active power, reactive power and rated capacity of the charging pile respectively.

5. The method for collaborative scheduling of virtual power plants based on aggregated distributed resources according to claim 4 is characterized in that: After collecting the real-time status information of resource points and issuing preliminary dispatch instructions, and before local processing and preliminary adjustment of resource point data, the response capacity is accurately quantified, including: Obtain the capacity declared by the user and verify it according to the capacity constraints; Obtain the actual corresponding capacity of adjacent time periods and determine whether the adjustment rate constraints are met; Based on the actual corresponding capacity, the corresponding time delay and the corresponding error are corrected, and the effective corresponding capacity is calculated.

6. The method for collaborative scheduling of virtual power plants based on aggregated distributed resources according to claim 1, characterized in that: The real-time status information of each resource point is collected through two-way communication technology, and preliminary dispatch instructions are issued, including: Using two-way communication technology, key status information of each distributed resource point is collected in real time; Pre-process the collected status information and monitor the operating status of each resource point in real time; Based on the pre-processed status information and historical data, the prediction algorithm is used to predict the energy demand in the future, and a preliminary scheduling strategy is formulated according to the prediction results and the real-time status of each resource point; The preliminary scheduling strategy is converted into specific scheduling instructions and sent to each distributed resource point through two-way communication technology.

7. The method for collaborative scheduling of virtual power plants based on aggregated distributed resources according to claim 1, characterized in that: The intelligent algorithm includes: Construct an optimization model for the operating cost of the virtual power plant system, where the operating cost is the minimum sum of the electricity purchase cost and the gas cost; Based on the optimization model, the electricity purchase cost and gas cost are calculated as follows: and Where i is the number of scheduling period periods, is the power purchased from the grid during period t, represents the electricity purchase price in time period t, i is the number of time periods in the dispatch cycle, is the power generation of the gas internal combustion engine during period t, c fc represents the natural gas price, F fuel is the calorific value of natural gas, η GE is the gas power generation efficiency, Δt is the time period length; The operating cost of the virtual power plant system is calculated based on the electricity purchase cost and gas cost. The formula is:

8. The method for collaborative scheduling of virtual power plants based on aggregated distributed resources according to claim 1, characterized in that: Based on edge computing technology, the data of each resource point is locally processed and preliminarily adjusted, and the processing results are uploaded to the cloud for global optimization, including: Preprocess the collected real-time operation data of distributed resources and calculate the status evaluation index based on voltage deviation, frequency deviation and power deviation; Determine whether to trigger a quick response based on the status assessment results. When the status indicator exceeds the preset threshold, select a solution from the preset control library for local preliminary adjustment. The processed data is compressed by extracting key feature parameters and using a time-space compression algorithm, and the compressed data is uploaded to the cloud in batches; Perform multivariate data alignment and fusion on the uploaded data of each edge node, establish a global state evaluation model, and build an optimization objective function based on the evaluation results to solve the global optimal solution; The optimized control strategy is sent to each edge node for state fine-tuning, and the execution result is fed back to the cloud.

9. A virtual power plant collaborative dispatching system based on aggregated distributed resources, characterized in that: The system comprises: The physical model building module aggregates distributed resources, builds a model based on physical characteristics, and quantifies the adjustable capacity and adjustment rate of each resource; The preliminary instruction issuing module collects the real-time status information of each resource point through two-way communication technology and issues preliminary dispatching instructions; The edge computing optimization module processes and preliminarily adjusts the data of each resource point locally based on edge computing technology, and uploads the processing results to the cloud for global optimization; The dispatch scheme acquisition module builds an optimized dispatch model based on the power system operating parameters, and dynamically adjusts the dispatch strategy through artificial intelligence algorithms to obtain a dynamic dispatch scheme; The scheme iteration optimization module obtains feedback data during the scheduling execution process, adjusts the parameters of the optimization scheduling model in real time, and iteratively optimizes the dynamic scheduling scheme.

10. The virtual power plant collaborative dispatching system based on aggregated distributed resources according to claim 9 is characterized in that: The edge computing optimization module includes: An evaluation index calculation unit pre-processes the collected real-time operation data of distributed resources and calculates the state evaluation index based on voltage deviation, frequency deviation and power deviation; The local preliminary adjustment unit determines whether to trigger a quick response based on the status evaluation results. When the status indicator exceeds the preset threshold, a solution is selected from the preset control library for local preliminary adjustment. The data compression upload unit compresses the processed data by extracting key feature parameters and using a time-space compression algorithm, and uploads the compressed data to the cloud in batches; The optimal solution solving unit performs multivariate data alignment and fusion on the uploaded data of each edge node, establishes a global state evaluation model, and constructs an optimization objective function based on the evaluation results to solve the global optimal solution; The strategy fine-tuning feedback unit sends the optimized control strategy to each edge node for state fine-tuning, and feeds back the execution result to the cloud.

Images