Optimal scheduling method of virtual power plant considering penetration of electric vehicles and flexible load

Abstract

The application relates to a virtual power plant optimal scheduling method considering electric vehicle penetration and flexible load, and belongs to the technical field of power system scheduling and new energy utilization, which comprises the following steps: collecting multi-element heterogeneous data and preprocessing; configuring a multi-electric vehicle penetration scenario and setting a new energy incentive weight coefficient; constructing an optimal scheduling model considering economy and low carbon and solving the model; generating device instructions and issuing the instructions under a differentiated strategy according to different penetration scenarios; real-time monitoring the actual operation state of each controllable device and iteratively optimizing the device to continuously improve the scheduling strategy. The application embeds the electric vehicle penetration as a core variable into the model, solves the poor adaptability of the scheduling strategy in different stages of traffic electrification, innovatively designs the new energy incentive weight coefficient, realizes the collaborative optimization of new energy and traditional energy, constructs a multi-type electric / thermal load collaborative model, fully excavates the flexibility of the load side, and optimizes the economic efficiency of power grid interaction through the collaborative charging and discharging of electric vehicles and energy storage.

Description

TECHNICAL FIELD

[0001] The present application relates to a virtual power plant optimization scheduling method considering electric vehicle penetration rate and flexible load, belonging to the technical field of power system dispatching and new energy utilization. BACKGROUND

[0002] New energy power generation represented by wind energy and solar energy has developed rapidly due to its cleanliness and sustainability, and its installed capacity and power generation proportion continue to increase, gradually replacing traditional fossil energy as the main power supply. However, new energy power generation has inherent intermittency and volatility, and large-scale distributed energy access significantly increases the complexity and uncertainty of power system operation; at the same time, distributed energy itself has problems such as small capacity, large quantity, unstable output, etc., and is difficult to independently participate in power trading, resulting in resource waste. Virtual power plant (VPP) is a unified agent for distributed resources to participate in power market. It aggregates dispersed distributed resources through a collaborative complementary mode to maximize overall efficiency. Flexible load has flexible scheduling characteristics and is effective in alleviating power supply and demand contradictions and improving new energy grid-connection efficiency. Some virtual power plants have included flexible load in scheduling to promote clean energy consumption such as wind and solar.

[0003] With the rapid development of electric vehicles (EV), their share in the automobile market continues to increase, and they have become an important part of power system load. Electric vehicle penetration rate, as a key indicator of the proportion of electric vehicles in the total number of vehicles, not only directly reflects the process of traffic electrification, but also has a profound impact on power system operation characteristics. By including electric vehicles as "mobile batteries" in virtual power plant scheduling, distributed energy consumption can be promoted, but existing research has the following shortcomings: (1) Most virtual power plant scheduling models do not consider the dynamic changes of electric vehicle penetration rate, ignoring the differentiated impact of electric vehicle cluster charging and discharging behavior on virtual power plant operation economy and grid stability under different penetration rates, resulting in models that cannot adapt to actual scenarios at different stages of development and distorted scheduling strategies; (2) The incentive mechanism for new energy generation is single, and it is difficult to achieve coordinated optimization of new energy and traditional energy through weight coefficients and other means, making it difficult to balance system economy and low carbon; (3) Load models focus on a single type of flexible load, and do not build a multi-type electric / thermal load collaborative scheduling mechanism with transferable, shiftable, and reducible characteristics, which cannot fully tap the flexibility potential on the load side.

[0004] Therefore, there is an urgent need for a virtual power plant optimization scheduling method that integrates "dynamic changes in electric vehicle penetration rate, multi-type flexible load coordination, and coordinated incentives for new energy and traditional energy" to reduce the total operating cost of the virtual power plant, improve new energy consumption rate, reduce dependence on traditional fossil energy, and promote the green and low-carbon transformation of the energy system. SUMMARY

[0005] In order to solve the above problems, the present application provides a virtual power plant optimal scheduling method considering electric vehicle penetration rate and flexible load.

[0006] The technical scheme adopted by the present application to solve the technical problems is:

[0007] In a first aspect, the present application provides a virtual power plant optimal scheduling method considering electric vehicle penetration rate and flexible load, comprising the following steps:

[0008] Step S1, collect multi-element heterogeneous data required for virtual power plant operation, and perform outlier processing, time scale alignment and data verification to form a standardized data set, wherein the multi-element heterogeneous data includes new energy output prediction data, user electricity / heat load prediction data, electric vehicle parameters and market environment data;

[0009] Step S2, based on the standardized data set, load the device cost coefficient and the flexible load compensation coefficient, and simultaneously configure multiple electric vehicle penetration rate scenarios according to the regional traffic electrification development level, and set the initial value of the new energy incentive weight coefficient;

[0010] Step S3, based on the initialized parameters and the configured scenarios, construct a virtual power plant optimal scheduling model with the minimum total operation cost of the virtual power plant as the objective function, and establish a complete constraint condition system, and obtain the optimal scheduling parameters by solving through an optimization algorithm, wherein the virtual power plant optimal scheduling model comprehensively considers the dual objectives of economy and low carbon;

[0011] Step S4, convert the obtained optimal scheduling parameters into specific instructions executable by each controllable device, and adopt differentiated charging and discharging strategies according to different electric vehicle penetration rate scenarios, and issue operation instructions to new energy units, traditional units, energy storage systems and electric vehicle clusters;

[0012] Step S5, real-time monitor the actual operation state of each controllable device, compare and analyze the monitoring data with the optimal scheduling parameters, take corresponding adjustment measures according to the deviation degree, and optimize the parameters based on the actual operation effect after each daily scheduling is completed, so as to realize the continuous improvement of the scheduling strategy.

[0013] Further, the step S1 comprises the following steps:

[0014] Step S11, synchronously collect the new energy output prediction curve and the traditional unit parameters on the energy side, the user electricity / heat load prediction curve and the flexible load characteristic parameters on the load side, the number of vehicles, the battery parameters and the user travel rules on the electric vehicle side, and the time-of-use electricity price curve and the carbon trading parameters on the market side;

[0015] Step S12, first identify and remove outliers from the collected raw data, then unify the data with different time resolutions to 1 hour time step, and finally perform data integrity check and reasonableness verification;

[0016] Step S13, integrate all data processed in step S12 into a structured standardized data set, which includes 1 hour step standardized data of energy side, load side, electric vehicle side, market and environment side.

[0017] Further, the step S2 comprises the following steps:

[0018] Step S21, based on the standardized data, load the new energy unit operation coefficient, traditional unit fuel coefficient, energy storage system operation coefficient and flexible load compensation coefficient;

[0019] Step S22, according to the national traffic electrification planning target and the regional electric vehicle development status, set five typical electric vehicle penetration rate levels of 0, 0.2, 0.3, 0.4 and 0.5;

[0020] Step S23, set the initial value range of new energy incentive weight coefficient α as 0.05-0.2, and through negative weight incentive new energy consumption, positive weight inhibits traditional energy consumption;

[0021] Step S24, form a structured "initialized parameter and configured scene" set.

[0022] Further, based on the five typical electric vehicle penetration rate levels of 0, 0.2, 0.3, 0.4 and 0.5 set in step S22, the charging and discharging strategy adjustment logic under different electric vehicle penetration rate scenarios is:

[0023] When the penetration rate is 0.2 in the low penetration rate scenario, the electric vehicle only charges in the 0:00-8:00 electricity price valley segment, and the SOC is maintained at 0.5-0.8, and does not participate in discharging in the peak period;

[0024] When the penetration rate is 0.3 / 0.4 in the medium-low / medium-high penetration rate scenario, the electric vehicle charges at full power in the 0:00-8:00 electricity price valley segment, and the SOC is raised to 0.7-0.9; In the 8:00-18:00 / 22:00-24:00 electricity price flat segment, when the SOC<0.6, charge; In the 18:00-22:00 electricity price peak segment, when the SOC>0.8, the electric vehicle discharges at a power of 20%-30% of the total capacity;

[0025] When the penetration rate is 0.5 in the high penetration rate scenario, the electric vehicle charges at full load in the valley segment, and the SOC is raised to 0.8-1.0, and discharges at a power of 30%-50% of the total capacity in the 8:00-22:00 electricity price peak segment.

[0026] Further, the step S3 comprises the following steps:

[0027] Step S31, a target function is constructed with the core of minimizing the total operation cost of virtual power plant in 24 hours, which comprehensively covers operation and maintenance cost, fuel cost, electricity purchase cost, electricity sale revenue, energy storage cost, flexible load compensation cost, carbon trading cost and electric vehicle charging and discharging cost / revenue;

[0028] Step S32, power balance constraint, heat balance constraint, equipment operation constraint, energy storage system constraint and electric vehicle charging and discharging constraint are established to form a complete constraint system;

[0029] Step S33, Cplex solver is called through Yalmip toolbox to solve the target function under the condition of meeting all constraints, and the optimal scheduling parameters in hours are obtained, and the optimal output plan of equipment and the optimal charging and discharging strategy of electric vehicle in each period are output.

[0030] Further, the target function is:

[0031] ,

[0032] ,

[0033] ,

[0034] ,

[0035] ,

[0036] ,

[0037] ,

[0038] ,

[0039] wherein, is the operation and maintenance cost of new energy unit, the new energy unit including wind turbine WT and photovoltaic PV; WT operation and maintenance coefficient and PV operation and maintenance coefficient respectively, are the output of WT and PV in t period respectively; is the fuel cost of traditional unit, the traditional unit including gas turbine GT and gas boiler GB; are GT and GB fuel coefficients respectively, are the output of GT and GB in t period respectively; is the electricity purchase cost; is the power purchased from the grid in t period, is the electricity purchase price in t period; for the electricity selling revenue; for the electricity selling power to the grid in time period t, for the electricity selling price in time period t; for the operation and maintenance cost of the energy storage system (ESS); for the operation and maintenance coefficient of the ESS, for the charging power of the ESS, for the discharging power of the ESS, for the heat storage power of the ESS; for the heat release power of the ESS; for the compensation cost of the flexible load; for the carbon trading cost; for the charging cost of the electric vehicle; for the discharging revenue of the electric vehicle; for the charging power of the electric vehicle in time period t, for the discharging power of the electric vehicle in time period t; for the new energy incentive term; for the fossil energy constraint term.

[0040] Further, the power balance constraint is:

[0041] ,

[0042] wherein, is the basic power load at time t; is the transferable power load at time t; is the shiftable power load at time t; is the reducible power load at time t;

[0043] The heat balance constraint is:

[0044] ,

[0045] wherein, is the gas turbine heat supply power (kW) at time t, is the gas boiler heat supply power (kW) at time t, is the basic heat load (kW) at time t, is the shiftable heat load (kW) at time t, is the transferable heat load (kW) at time t, is the reducible heat load (kW) at time t;

[0046] The equipment operation constraint (unit power upper and lower limit constraint) is:

[0047] ,

[0048] wherein, For the rated power of the gas turbine, For the maximum output of the WT, For the maximum output of the GB, For the maximum output of the PV, For the minimum value of the power exchanged with the grid, For the maximum value of the power exchanged with the grid.

[0049] The energy storage system constraint is;

[0050] ,

[0051] Wherein, The minimum value of the energy storage charging / discharging power is, The maximum value of the energy storage charging / discharging power is, The minimum value of the energy storage heat storage / dissipation power is, The maximum value of the energy storage heat storage / dissipation power is, The state of charge of the energy storage is, The state of discharge of the energy storage (0-1 variable, 1 indicates operation and 0 indicates stop), and charging and discharging are not performed at the same time;

[0052] The electric vehicle charging and discharging constraint is:

[0053] ,

[0054] Wherein, The maximum value of the ESS charging power is, The maximum value of the ESS discharging power is.

[0055] Further, in the step S31, the new energy incentive weight coefficient α introduced is used to realize the collaborative optimization of the economic and low-carbon objectives through the following mechanism:

[0056] When the system new energy output is sufficient, the negative weight value of the new energy incentive weight coefficient α is increased, the new energy output is converted into system cost deduction, and the new energy consumption is encouraged;

[0057] When the system power supply is tight, the positive weight value of the new energy incentive weight coefficient α is increased, the cost of traditional high-carbon equipment is highlighted, and the excessive consumption of fossil energy is inhibited.

[0058] Further, the step S4 comprises the following steps:

[0059] Step S41, the obtained optimal scheduling parameters in hours are converted into control signals recognizable by the device controller, and the control signals comprise new energy unit output instructions, traditional unit start-stop instructions, energy storage system charging and discharging instructions, and electric vehicle cluster charging and discharging strategies;

[0060] Step S42, different charging and discharging strategies are adopted according to different electric vehicle penetration rate scenarios, and the low penetration rate scenario gives priority to guaranteeing the user travel demand, and the high penetration rate scenario fully plays the mobile energy storage characteristics of the electric vehicle;

[0061] Step S43, the operation instruction is issued to each distributed resource through the virtual power plant central control system, and the effective execution of the scheduling plan is ensured.

[0062] Further, the step S5 comprises the following steps:

[0063] Step S51, the actual operation data of each device is collected in real time through the SCADA system, and the actual operation data includes the actual output of new energy, the actual output of traditional units, the actual state of energy storage system and the actual charging and discharging behavior of electric vehicles;

[0064] Step S52, the actual operation data is compared with the obtained optimal scheduling parameter, the deviation degree is calculated and the deviation reason is analyzed;

[0065] Step S53, according to the deviation analysis result, when the deviation is less than 5%, the original instruction is maintained, and when the deviation is greater than or equal to 5%, the model solving step is re-executed to generate a new instruction;

[0066] Step S54, after the daily scheduling is completed, the new energy incentive weight coefficient and the electric vehicle charging and discharging strategy are updated based on the actual operation effect, and the continuous optimization of the scheduling model is realized.

[0067] The beneficial effects of the technical scheme of the embodiment of the application are as follows:

[0068] The virtual power plant optimal scheduling method considering electric vehicle penetration rate and flexible load of the technical scheme of the embodiment of the application comprises the following steps: collecting multi-element heterogeneous data and preprocessing; configuring multiple electric vehicle penetration rate scenarios and setting new energy incentive weight coefficients; constructing an optimal scheduling model considering economy and low carbon and solving; generating device instructions and issuing them according to different penetration rate scenarios using differentiated strategies; real-time monitoring of the actual operation state of each controllable device and iterative optimization to continuously improve the scheduling strategy. The application embeds the electric vehicle penetration rate as a core variable into the model, solves the poor adaptability of the scheduling strategy in different stages of traffic electrification, innovatively designs the new energy incentive weight coefficient to realize the collaborative optimization of new energy and traditional energy, constructs a multi-type electric / thermal load collaborative model to fully tap the flexibility of the load side, and optimizes the economic efficiency of the power grid interaction through the collaborative charging and discharging of electric vehicles and energy storage. Experiments show that the application can effectively reduce the operation cost, improve the new energy consumption rate, and promote the low-carbon transformation of the power system.

[0069] This invention embeds electric vehicle penetration rate into a virtual power plant optimization model. By setting up multiple penetration rate scenarios and combining a Cplex solver to quantify scheduling parameters (such as renewable energy output, traditional generator output, and charging / discharging power) under different penetration rates, the virtual power plant optimization model can adapt to different stages of transportation electrification development. This solves the technical problem of poor adaptability of existing virtual power plant scheduling strategies under different electric vehicle penetration rates, achieving dynamic optimization across multiple scenarios and avoiding scheduling strategy distortion. Experiments show that when the penetration rate increases to 50%, the total operating cost of the virtual power plant can be continuously reduced, while renewable energy output increases by 9.2% and gas turbine output decreases by 15.4%, effectively adapting to the scheduling needs of different development stages and avoiding the "one-size-fits-all" defects of traditional models.

[0070] This invention constructs a two-way adjustment mechanism through the innovative design of a "new energy incentive weighting coefficient α": on the one hand, it uses a negative weight to convert the output of new energy (wind power, photovoltaic) into system cost deduction, actively tapping the potential for new energy consumption; on the other hand, it uses a positive weight to highlight the cost of traditional high-carbon equipment (gas boilers), suppressing excessive consumption of fossil fuels. This mechanism breaks through the limitations of traditional single-objective scheduling, achieving synergistic optimization of new energy and traditional energy, both increasing the participation of new energy and reducing carbon emissions, promoting the green and low-carbon transformation of virtual power plants. This synergistic effect cannot be achieved by existing technologies that only focus on economic costs or single new energy consumption.

[0071] This invention constructs a multi-type electric / thermal load coordination model that is "shiftable, transferable, and reduceable," guiding load optimization through differentiated compensation coefficients: shiftable loads enable time-based transfers, transferable loads achieve spatiotemporal dual-dimensional adjustments, and reduceable loads address emergency scenarios. Simultaneously, it establishes power balance and thermal balance constraints, coordinating the electric / thermal output of gas turbines / boilers with energy storage and the electric / thermal storage of electric vehicles (e.g., supplementing electricity through electric vehicle discharge and heat through gas boiler heating during peak hours). This invention reduces the peak-valley load difference, improves the renewable energy absorption rate, and achieves coupled optimization of the electric / thermal system, far exceeding the scheduling effects of existing single-load-type and single-energy-form methods.

[0072] This invention utilizes the coordinated charging and discharging of electric vehicles and energy storage systems: during off-peak hours (0:00-8:00), the system controls the charging and storage of electric vehicles and energy storage to store electrical energy; during peak hours (18:00-22:00), the system uses the discharge of both systems to supplement the system's power demand, reducing the need for high-priced electricity purchases; simultaneously, excess renewable energy is fed back to the grid to generate revenue from electricity sales. Experimental data shows that after introducing the electric vehicle dispatch model, the grid's electricity purchase cost is reduced by up to 6.7%, and the revenue from electric vehicle discharge can offset some of the charging costs, significantly optimizing the economic interaction between the virtual power plant and the grid. This cost optimization effect is unattainable by existing dispatch technologies that do not integrate the charging and discharging characteristics of electric vehicles. Attached Figure Description

[0073] Figure 1 is a flow chart of a virtual power plant optimization scheduling method considering electric vehicle penetration rate and flexible load according to an example embodiment;

[0074] Figure 2 is a schematic diagram of a virtual power plant basic framework according to an example embodiment;

[0075] Figure 3 is a specific implementation flow chart of realizing virtual power plant optimization scheduling according to an example embodiment;

[0076] Figure 4 is a time-of-use electricity purchase and sale price curve diagram according to an example embodiment. DETAILED DESCRIPTION

[0077] In order to more clearly illustrate the technical features of the scheme of the present application, the present application will be described in detail below with reference to specific embodiments and accompanying drawings.

[0078] As shown in Figure 1 , the virtual power plant optimization scheduling method considering electric vehicle penetration rate and flexible load provided by the embodiment of the present application comprises the following steps:

[0079] Step S1, collect multi-element heterogeneous data required for virtual power plant operation, and perform outlier processing, time scale alignment and data verification to form a standardized data set, wherein the multi-element heterogeneous data comprises new energy output prediction data, user electricity / heat load prediction data, electric vehicle parameters and market environment data;

[0080] Step S2, based on the standardized data set, load device cost coefficients and flexible load compensation coefficients, and simultaneously configure multiple electric vehicle penetration rate scenarios according to the regional traffic electrification development level, and set the initial value of the new energy incentive weight coefficient;

[0081] Step S3, based on the initialized parameters and the configured scenarios, construct a virtual power plant optimization scheduling model with the minimization of the total operation cost of the virtual power plant as the objective function, and establish a complete constraint condition system, and obtain the optimal scheduling parameters by solving through an optimization algorithm, wherein the virtual power plant optimization scheduling model comprehensively considers the dual objectives of economy and low carbon;

[0082] Step S4, convert the obtained optimal scheduling parameters into specific instructions executable by each controllable device, and adopt differentiated charging and discharging strategies according to different electric vehicle penetration rate scenarios, and issue operation instructions to new energy units, traditional units, energy storage systems and electric vehicle clusters;

[0083] Step S5, real-time monitoring of the actual running state of each controllable device, comparing and analyzing the monitoring data with the optimal scheduling parameters, taking corresponding adjustment measures according to the deviation degree, and optimizing the parameters based on the actual running effect after the daily scheduling is completed, so as to realize the continuous improvement of the scheduling strategy.

[0084] Further, the step S1 comprises the following steps:

[0085] Step S11, synchronously collecting the new energy output prediction curve of the energy side and the traditional unit parameter, the user electricity / heat load prediction curve of the load side and the flexible load characteristic parameter, the retention quantity, the battery parameter and the user travel rule of the electric vehicle side, and the time-of-use electricity price curve and the carbon trading parameter of the market side;

[0086] Step S12, first, identifying and eliminating the outliers of the collected original data, then unifying the data of different time resolutions to 1 hour time step, and finally, performing data integrity check and reasonableness verification;

[0087] Step S13, integrating all the data processed in step S12 into a structured standardized data set, wherein the standardized data set comprises 1 hour step standardized data of the energy side, the load side, the electric vehicle side, the market and the environment side.

[0088] Further, the step S2 comprises the following steps:

[0089] Step S21, based on the standardized data, loading the new energy unit operation coefficient, the traditional unit fuel coefficient, the energy storage system operation coefficient and the flexible load compensation coefficient;

[0090] Step S22, according to the national traffic electrification planning target and the regional electric vehicle development status, setting five typical electric vehicle penetration rate levels of 0, 0.2, 0.3, 0.4 and 0.5;

[0091] Step S23, setting the initial value range of the new energy incentive weight coefficient α as 0.05-0.2, and through the negative weight incentive, the new energy consumption is encouraged, and the traditional energy consumption is inhibited;

[0092] Step S24, forming a structured "initialized parameter and configured scene" set.

[0093] Further, based on the five typical electric vehicle penetration rate levels of 0, 0.2, 0.3, 0.4 and 0.5 set in step S22, the charging and discharging strategy adjustment logic under different electric vehicle penetration rate scenes is:

[0094] When the penetration rate is 0.2 in the low penetration rate scene, the electric vehicle only charges in the 0:00-8:00 electricity price valley segment, the SOC is maintained at 0.5-0.8, and does not participate in discharging in the peak period.

[0095] When the penetration rate is 0.3 / 0.4, the electric vehicle charges at full power in the valley period of 0:00-8:00, and the SOC is raised to 0.7-0.9; when the SOC is less than 0.6, the electric vehicle charges in the flat period of 8:00-18:00 / 22:00-24:00; when the SOC is greater than 0.8, the electric vehicle discharges at a power of 20%-30% of the total capacity in the peak period of 18:00-22:00;

[0096] When the penetration rate is 0.5, the electric vehicle charges at full power in the valley period, and the SOC is raised to 0.8-1.0; the electric vehicle discharges at a power of 30%-50% of the total capacity in the peak period of 8:00-22:00.

[0097] Further, the step S3 comprises the following steps:

[0098] Step S31, a target function is constructed with the minimization of the 24-hour total operation cost of the virtual power plant as the core, which comprehensively covers the operation and maintenance cost, fuel cost, electricity purchase cost, electricity sale income, energy storage cost, flexible load compensation cost, carbon trading cost and electric vehicle charging and discharging cost / income;

[0099] Step S32, power balance constraints, heat balance constraints, device operation constraints, energy storage system constraints and electric vehicle charging and discharging constraints are established to form a complete constraint system;

[0100] Step S33, the Cplex solver is called through the Yalmip toolbox to solve the target function under the condition of meeting all the constraints, and the optimal scheduling parameters in hours are obtained, and the optimal output plan of each period of device and the optimal charging and discharging strategy of electric vehicle are output.

[0101] Further, the target function is:

[0102] ,

[0103] ,

[0104] ,

[0105] ,

[0106] ,

[0107] ,

[0108] ,

[0109] ,

[0110] in, The cost of operation and maintenance of new energy units, which include wind power (WT) and photovoltaic (PV); These are the WT maintenance coefficient and the PV maintenance coefficient, respectively. These represent the outputs of WT and PV during time period t, respectively. The fuel cost of conventional generating units, which include gas turbines (GT) and gas-fired boilers (GB); These are the GT and GB fuel coefficients, respectively. These represent the outputs of GT and GB during time period t, respectively. For electricity purchase costs; Let t be the power purchased from the grid during time period t. The electricity purchase price for period t; For revenue from electricity sales; Let t be the amount of electricity sold to the grid during time period t. The electricity price for period t; For the operation and maintenance costs of the energy storage system (ESS); This is the ESS operation and maintenance coefficient. Charging power for ESS For ESS discharge power, For ESS thermal storage capacity; For ESS heat dissipation power; To compensate for costs related to flexible loads; For carbon trading costs; The cost of charging electric vehicles; For the discharge revenue of electric vehicles; The charging power of electric vehicles during time period t. The discharge power of the electric vehicle during time period t; For new energy incentive items; This is a fossil fuel constraint.

[0111] Furthermore, the power balance constraint is:

[0112] ,

[0113] in, Let t be the basic electrical load; The electrical load can be transferred at time t; The electrical load can be shifted at time t; The electrical load can be reduced at time t;

[0114] The thermodynamic balance constraint is:

[0115] ,

[0116] in, is the heat supply power of the gas turbine at time t (kW), is the heat supply power of the gas boiler at time t (kW), is the basic thermal load at time t (kW), is the translatable thermal load at time t (kW), is the transferred thermal load at time t (kW), is the reducible thermal load at time t (kW);

[0117] The equipment operation constraints (unit power upper and lower limit constraints) are:

[0118] ,

[0119] wherein, is the rated power of the gas turbine, is the WT predicted maximum output, is the GB predicted maximum output, is the PV predicted maximum output, is the minimum value of the power exchanged with the grid, is the maximum value of the power exchanged with the grid;

[0120] The energy storage system constraints are:

[0121] ,

[0122] wherein, is the minimum value of the energy storage charge / discharge power, are the maximum values of the energy storage charge / discharge power, is the minimum value of the energy storage heat storage / heat release power, is the maximum value of the energy storage heat storage / heat release power, is the state of charge of the energy storage, is the state of discharge of the energy storage (0-1 variable, 1 indicates operation and 0 indicates stop), and charge and discharge are not performed at the same time;

[0123] The electric vehicle charging and discharging constraints are:

[0124] ,

[0125] wherein, is the maximum value of the ESS charging power, is the maximum value of the ESS discharging power.

[0126] Further, in the step S31, the new energy incentive weight coefficient α introduced realizes the synergistic optimization of the economic and low-carbon objectives through the following mechanism:

[0127] When the system new energy output is sufficient, the negative weight value of the new energy incentive weight coefficient alpha is increased, the new energy output is converted into system cost deduction, and the new energy consumption is encouraged;

[0128] When the system power supply is tight, the positive weight value of the new energy incentive weight coefficient alpha is increased, the cost of traditional high-carbon equipment is highlighted, and the excessive consumption of fossil energy is inhibited.

[0129] The application adopts an electric vehicle penetration rate dynamic integration technology, and the electric vehicle penetration rate As a core variable embedded in the scheduling model, multiple penetration rate scenarios such as 0, 0.2, 0.3, 0.4 and 0.5 are set, the influence of the charging and discharging behavior of the electric vehicle cluster on the operation of the virtual power plant under different penetration rates is quantified, the Cplex solver is called through the Yalmip toolbox to realize efficient solving of the scheduling model under multiple scenarios, and key parameters such as new energy output, traditional unit output, energy storage and electric vehicle charging and discharging power at each time are output, thereby supporting dynamic scheduling decision.

[0130] Further, the step S4 comprises the following steps:

[0131] Step S41, the obtained optimal scheduling parameter in hours is converted into a control signal recognizable by a device controller, and the control signal comprises a new energy unit output instruction, a traditional unit start-stop instruction, a storage system charging and discharging instruction and an electric vehicle cluster charging and discharging strategy;

[0132] Step S42, different charging and discharging strategies are adopted according to different electric vehicle penetration rate scenarios, the user travel demand is preferentially guaranteed in a low penetration rate scenario, and the mobile energy storage characteristics of the electric vehicle are fully utilized in a high penetration rate scenario;

[0133] Step S43, the operation instruction is issued to each distributed resource through the virtual power plant central control system, and the effective execution of the scheduling plan is ensured.

[0134] Further, the step S5 comprises the following steps:

[0135] Step S51, actual operation data of each device are collected in real time through the SCADA system, and the actual operation data comprises actual new energy output, actual traditional unit output, actual storage system state and actual electric vehicle charging and discharging behavior;

[0136] Step S52, the actual operation data are compared with the obtained optimal scheduling parameter, the deviation degree is calculated and the deviation reason is analyzed;

[0137] Step S53, according to the deviation analysis result, when the deviation is less than 5%, the original instruction is maintained, and when the deviation is greater than or equal to 5%, a new instruction is generated by re-executing the model solving step;

[0138] Step S54, after the daily dispatch is completed, the new energy incentive weight coefficient and the electric vehicle charging and discharging strategy are updated based on the actual operation effect, and the continuous optimization of the dispatching model is realized.

[0139] As shown in Figure 2 The virtual power plant system architecture of the application is built by using multi-dimensional system architecture construction technology, which includes "new energy power generation unit (wind turbine WT, photovoltaic unit PV) + traditional power generation unit (gas turbine, gas boiler) + energy storage system (ESS) + electric vehicle cluster + multi-type user load", the user load is subdivided into basic load, translatable load, transferable load and reducible load according to demand response characteristics, the classification control and collaborative scheduling of electric / thermal load are realized, and the flexibility potential of the load side is fully tapped.

[0140] The virtual power plant system architecture includes:

[0141] Energy supply unit: the virtual power plant system includes new energy power generation unit (wind turbine WT, photovoltaic unit PV), traditional power generation unit (gas turbine, gas boiler);

[0142] Energy storage unit: energy storage system (ESS), electric vehicle cluster (as mobile energy storage);

[0143] Load unit: electric load, user load;

[0144] The user load is divided into four categories based on demand response participation:

[0145] Basic load: meets basic electricity / heat demand, cannot be adjusted or interrupted, and has extremely high requirements for power supply reliability;

[0146] Translatable load: total power consumption / heat consumption is fixed, and power consumption / heat consumption time can be flexibly adjusted within a certain period of time;

[0147] Transferable load: both power consumption / heat consumption time can be adjusted and load transfer in geographical space can be realized;

[0148] Reducible load: can reduce or suspend operation in emergency scenarios such as power grid failure, and can quickly respond to system dispatching requirements.

[0149] Dispatching control unit: Cplex solver based on Yalmip toolbox in Matlab 2022a environment, performs optimization calculation and instruction issuance.

[0150] As shown in Figure 3 The application realizes virtual power plant optimal dispatching through the process of data acquisition and preprocessing, parameter initialization and scene configuration, multi-objective model construction and solution, dispatching instruction generation and issuance, and running monitoring and result iteration optimization. The specific implementation process is as follows.

[0151] I. Data Acquisition and Preprocessing (Input: Raw data; Output: Standardized dataset).

[0152] Data collection:

[0153] Energy-side data: 24-hour power output forecast curves for renewable energy (WT / PV) (generated from historical power data and meteorological forecast models), rated power, efficiency curves, and fuel consumption coefficients of conventional units (refer to equipment manuals); charging and discharging efficiency of energy storage systems (ESS), and upper limits of electrical / thermal storage capacity;

[0154] Load-side data: User electricity / heat load forecast curves (based on historical load data + user behavior models, such as peak residential load period 18:00-22:00), adjustment range and response delay of flexible loads (which can be shifted / transferred / reduced) (determined through user demand surveys, such as shiftable load adjustment window ±2 hours).

[0155] Electric vehicle side data: total number of vehicles in the region, electric vehicle ownership, electric vehicle battery capacity, charging and discharging efficiency, and user travel patterns.

[0156] Market and environmental data: 24-hour time-of-use electricity price curve (e.g.) Figure 4 As shown, referencing the peak-valley electricity prices published by the power grid company (e.g., 0.3 yuan / kWh for electricity purchase during the off-peak period from 0:00 to 8:00), carbon trading prices, and equipment carbon emission coefficients.

[0157] Data preprocessing:

[0158] Outlier handling: Remove extreme values ​​from renewable energy output and load forecasts;

[0159] Data alignment: Unify all data to a 1-hour time step (matching the scheduling cycle), such as aggregating electric vehicle charging data at 15-minute intervals into hourly average power.

[0160] Data verification: Verify the rationality of the upper limit of energy output and the demand of load side (e.g., the predicted output of WT does not exceed the rated power, and the basic load is not less than 0).

[0161] This step provides accurate data support for subsequent parameter initialization and model building, avoiding the distortion of scheduling strategies caused by abnormal data.

[0162] II. Parameter Initialization and Scene Configuration (Input: Standardized dataset; Output: Initialization parameters + scene scheme).

[0163] Basic parameter initialization (directly linked to data acquisition results to ensure parameter authenticity): Loading WT / PV operation and maintenance coefficients, gas turbine / boiler fuel coefficients: WT operation and maintenance coefficients = 0.72 yuan / kW, PV operation coefficient = 0.52 yuan / kW, gas turbine / boiler fuel coefficient = 2.5 yuan / kW, ESS operation coefficient = 0.5 yuan / k;

[0164] Set flexible load compensation coefficient: translatable load compensation coefficient = 0.2 yuan / kW, transferable load compensation coefficient = 0.3 yuan / kW, reducible load compensation coefficient = 0.4 yuan / k;

[0165] Configure different electric vehicle penetration rate scenarios: electric vehicle penetration rate (K e ᵥ) represents the proportion of the number of electric vehicles to the total number of vehicles in the region, and the value range is 0~0.5;

[0166] Optimize the new energy incentive weight coefficient α, and adjust by negatively weighting the new energy income and positively weighting the fossil energy cost. The cost and new energy consumption rate under different penetration rates are shown in Table 1.

[0167] Table 1 Comparison of cost and new energy consumption rate under different penetration rates

[0168]

[0169] The basis for the scenario division of the application is as follows: combined with the national / regional traffic electrification plan (such as "New Energy Vehicle Industry Development Plan (2021-2035)", which proposes that the electric vehicle penetration rate will reach 20% in 2025), the retail penetration rate of new energy vehicles in China reaches 55.3%, and the current electric vehicle popularization status in the region, five typical scenarios are divided: K ev = 0 (no electric vehicle), 0.2 (low penetration), 0.3 (medium-low penetration), 0.4 (medium-high penetration), and 0.5 (high penetration).

[0170] This step determines the model core parameters and the electric vehicle penetration rate scenario, and provides boundary conditions for the multi-objective optimization model.

[0171] Three, multi-objective model construction and solution (input: initialization parameters + scenario scheme; output: optimal scheduling parameters).

[0172] 3.1 Take the minimum total operation cost of the virtual power plant in 24 hours as the objective function, while meeting the power balance, heat balance, upper and lower limits of each unit power, energy storage system operation, electric vehicle charging and discharging and other constraint conditions. The objective function is as follows:

[0173] ,

[0174] ,

[0175] ,

[0176] ,

[0177] ,

[0178] ,

[0179] ,

[0180] ,

[0181] wherein, is the operation and maintenance cost of new energy units, the new energy units including wind turbines WT and photovoltaic PV; are the WT operation and maintenance coefficient and the PV operation and maintenance coefficient respectively, = 0.72 yuan / kW, = 0.52 yuan / kW, are the output (kW) of the WT and the PV in the t period respectively; is the fuel cost of traditional units, the traditional units including gas turbines GT and gas boilers GB; are the GT and GB fuel coefficients respectively, = 2.5 yuan / kW, are the output (kW) of the GT and the GB in the t period respectively; is the electricity purchase cost; is the electricity purchase power (kW) in the t period, is the electricity purchase price in the t period; is the electricity sale revenue; is the electricity sale power (kW) to the grid in the t period, is the electricity sale price in the t period; is the operation and maintenance cost of the energy storage system ESS; is the ESS operation and maintenance coefficient, = 0.5 yuan / kW, is the ESS charging power, is the ESS discharging power, is the ESS heat storage power; is the ESS heat release power; is the flexible load compensation cost; is the carbon trading cost; is the electric vehicle charging cost; is the electric vehicle discharging revenue; is the electric vehicle charging power in the t period, is the electric vehicle discharging power in the t period; New energy incentive term; Fossil energy constraint term.

[0182] Flexible load compensation cost The compensation fee paid by the virtual power plant for guiding the user to adjust the flexible load (translatable, transferable, and reducible load) is composed of three parts:

[0183] ,

[0184] Translatable load compensation: ,

[0185] Transferable load compensation: ,

[0186] Reducible load compensation: ,

[0187] Carbon trading cost :

[0188] ,

[0189] Wherein, is the total carbon emission (g), is the carbon quota (g), and U is the device set, is the device birth stage carbon emission coefficient, is the device running stage carbon emission coefficient;

[0190] New energy incentive term : ,

[0191] Fossil energy constraint term : .

[0192] The present application takes the minimization of the total operation cost of the virtual power plant in 24 hours as the core, covers the operation and maintenance cost, fuel cost, purchase / sale electricity cost, energy storage operation and maintenance cost, flexible load compensation cost, carbon trading cost, electric vehicle charging and discharging cost / revenue, simultaneously introduces a new energy incentive weight coefficient α, and converts the new energy output into system cost deduction through a negative weight (α < 0) Through a positive weight, the cost of traditional high-carbon equipment is highlighted (α > 0) , and the economic and low-carbon targets are synergistically optimized.

[0193] The new energy incentive weight coefficient a is obtained according to the solving result, in order to realize the collaborative promotion of new energy consumption rate and traditional energy substitution rate, and ensure that the virtual power plant meets the low-carbon economic operation requirement under different penetration rates. The new energy incentive weight coefficient a is between 0.05-0.2, which can not only ensure the effective landing of the bidirectional regulation function of the new energy incentive weight coefficient a, but also maintain the balance between the economic cost and the low-carbon target of the virtual power plant. If the new energy incentive weight coefficient a is too small, it cannot play the role of stimulating new energy and reducing the consumption of fossil energy. If the new energy incentive weight coefficient a is too large, it will lead to unstable power supply in peak period. VPP needs to rely on electric vehicle discharge and energy storage discharge to supplement energy at 18:00-22:00 (late peak), at this time, too much constraint on gas turbine output may lead to a power supply gap of "insufficient new energy output + limited fossil energy + limited electric vehicle energy storage" in the late peak period, which needs to be purchased from the power grid to increase the cost.

[0194] 3.2 Build constraint conditions: build a complete constraint system including power balance constraints (ensure supply and demand matching), heat balance constraints (adapt to electricity / heat cooperation), unit power upper and lower limit constraints (safe operation of equipment), energy storage charge and discharge exclusion constraints (avoid equipment damage), electric vehicle SOC and charge and discharge state constraints (ensure user vehicle demand).

[0195] The power balance constraint is:

[0196] ,

[0197] Among them, is the basic electric power load (kW) at time t; is the transferable electric power load at time t; is the translatable electric power load at time t; is the reducible electric load at time t;

[0198] The heat balance constraint is:

[0199] ,

[0200] Among them, is the gas turbine heat supply power (kW) at time t, is the gas boiler heat supply power (kW) at time t, is the basic heat load (kW) at time t, is the translatable heat load (kW) at time t, is the transferable heat load (kW) at time t, is the reducible heat load (kW) at time t;

[0201] The equipment operation constraint (unit power upper and lower limit constraint) is:

[0202] ,

[0203] where, is the gas turbine rated power (kW), is the WT predicted maximum output (kW), is the GB predicted maximum output (kW), is the PV predicted maximum output (kW), is the minimum value of the power exchanged with the grid (kW), is the maximum value of the power exchanged with the grid (kW);

[0204] The energy storage system constraints are:

[0205] ,

[0206] where, is the minimum value of the energy storage charge / discharge power, are the maximum values of the energy storage charge / discharge power, is the minimum value of the energy storage heat storage / heat release power (kW), is the maximum value of the energy storage heat storage / heat release power (kW), is the energy storage state of charge, is the energy storage state of discharge (0-1 variable, 1 means running, 0 means stopping), and charge and discharge do not occur at the same time;

[0207] The electric vehicle charging and discharging constraints are:

[0208] ,

[0209] where, is the maximum value of the ESS charging power (kW), is the maximum value of the ESS discharging power (kW).

[0210] The derivation process of the thermal balance constraint (constraint condition of the scheduling of the adaptation of the electric / thermal synergy) is as follows:

[0211] The gas turbine electric-thermal coupling relationship derivation (based on energy conservation and equipment efficiency):

[0212] Given: the fuel input heat Q in (kJ) of the gas turbine, the power generation efficiency η e =P GT / Q in (P GT is the power generation power of the gas turbine, kW), the heat supply efficiency η h =Q GT / Q in (Q GT is the heat supply power of the gas turbine, kW).

[0213] Derivation: From η e =P GT / Q in →Q in =P GT / η e Substitute η h =Q GT / Q in →Q GT =P GT ×η h / η e .

[0214] Derivation of thermal equilibrium constraints (based on "supply = demand + storage change"):

[0215] Energy supply side: Gas turbine heating Q GT (t) + Gas-fired boiler heating Q GB (t);

[0216] Energy consumption side: Basic heat load Q base (t) + Transferable heat load Q shift (t) + Transferable heat load Q trans (t) + can reduce heat load Q cut (t);

[0217] The equilibrium equation is obtained: Q GT (t)+Q GB (t)=Q base (t)+Q shift (t)+Q trans (t)+Q cut (t).

[0218] The establishment of a full-scenario constraint system transforms scheduling requirements into mathematical models, and outputs the operating parameters of each device through a solver, ensuring the feasibility and security of the scheduling strategy.

[0219] IV. Generation and Issuance of Scheduling Instructions (Input: Optimal Scheduling Parameters; Output: Equipment Operation Instructions).

[0220] The present invention first converts the scheduling parameters, which are based on an hourly time scale (for example, the output of wind power (WT) is 120kW in the first hour), into control signals that are adapted to different equipment.

[0221] For new energy units, taking wind power and photovoltaic (WT / PV) as examples, power commands at specific times are directly issued to their controllers to ensure that the output of new energy meets the dispatch requirements;

[0222] For the electric vehicle (EV) cluster, send instructions containing charging power and battery state of charge (SOC) targets to the electric vehicle aggregator, which can coordinate the charging and discharging behavior of electric vehicles to meet user needs and provide auxiliary services for the power grid.

[0223] This step converts the solution into device executable scheduling instructions. The implementation steps of the scheduling are as follows:

[0224] Data collection: Obtain new energy (WT, PV) output prediction curves, user electricity / heat load prediction curves, time-of-use electricity price curves, electric vehicle parameters (battery capacity, charging and discharging efficiency, SOC upper and lower limits), and various cost coefficients (operation and maintenance, fuel, compensation, carbon trading) within 24 hours;

[0225] Parameter initialization: Load device cost coefficients, flexible load compensation coefficients, electric vehicle parameters, weight coefficients, etc.

[0226] Model solving: Optimize the objective function through the Cplex solver, and output the output of each device, electric vehicle charging and discharging power, and flexible load adjustment amount at time t;

[0227] Instruction issuance: The scheduling control unit issues operation instructions to WT, PV, GT, ESS, and EV cluster.

[0228] Five, operation monitoring and result iterative optimization (input: device operation feedback data; output: optimized parameters), correct scheduling deviation, and improve model adaptability.

[0229] According to the model solving result, the new energy incentive weight coefficient α is adjusted to realize the coordinated improvement of new energy consumption rate and traditional energy replacement rate, and to ensure that the virtual power plant meets the low-carbon economic operation requirements under different penetration rates. The specific implementation process is as follows:

[0230] Real-time data monitoring: Collect actual operation data of each device (such as WT actual output, EV actual SOC) through the SCADA system, compare with the optimal scheduling parameters, and calculate the deviation (such as = actual output - predicted output);

[0231] Deviation judgment and processing:

[0232] If the deviation is less than 5% (normal fluctuation): maintain the current instruction;

[0233] If the deviation is greater than or equal to 5% (such as sudden drop in wind power output): trigger emergency optimization, re-execute "model solving" (based on the actual new energy consumption rate and system supply and demand deviation of each day, adjust the value of α using proportional-integral (PI) controller or lookup table method, increase the output of gas turbine), generate new instructions and issue.

[0234] Daily iterative optimization: After the daily scheduling is completed, the actual operation data (such as total cost and new energy consumption) are used to update the new energy incentive weight coefficient α value and the electric vehicle charging and discharging strategy.

[0235] Based on the balance of the "user demand-system supply and demand" double target, the electric vehicle charging and discharging strategy adjustment logic of the application in different scenarios is:

[0236] K ev =0.2 (low penetration): priority is given to meeting user travel demand, and the electric vehicle only charges in the valley section (0:00-8:00) of the electricity price, and the SOC is maintained at 0.5-0.8 (to ensure next day commuting), and no discharging is performed in the peak period (18:00-22:00) (to avoid affecting user vehicle use), and the time-of-use purchase and sale electricity price is as shown in Figure 4

[0237] K ev =0.5 (high penetration): the electric vehicle participates in peak shaving as "mobile energy storage", and is fully charged in the valley section (0:00-8:00) (SOC is raised to 0.8-1.0), and is discharged to supplement system power in the peak section (18:00-22:00) (discharging power accounts for 30%-50% of the total capacity of the electric vehicle), and is flexibly adjusted according to the SOC in the flat section (8:00-18:00 / 22:00-24:00) (a small amount of charging is performed when the SOC is lower than 0.5).

[0238] Compared with the prior art, the application has the following characteristics:

[0239] 1. The technical problem of "poor adaptability of virtual power plant scheduling strategy under different electric vehicle penetration rates" is solved, and multi-scenario dynamic optimization is achieved. The electric vehicle penetration rate is not taken as a core variable in the scheduling model in the prior art, which cannot adapt to actual scenarios in different stages of traffic electrification (low / medium / high penetration rate), resulting in distorted scheduling strategy. The technical solution of the application embeds the electric vehicle penetration rate in the virtual power plant optimization model, sets multiple penetration rate scenarios, and quantifies scheduling parameters (such as new energy output, traditional unit output, and charging and discharging power) under different penetration rates by using a Cplex solver. Experiments show that when the penetration rate is increased to 50%, the total operation cost of the virtual power plant can be continuously reduced, the new energy output is increased by 9.2%, and the gas turbine output is decreased by 15.4%, effectively adapting to the scheduling requirements of different development stages and avoiding the defects of "one-size-fits-all" of traditional models.

[0240] ​2. This solution addresses the technical challenge of insufficient synergistic optimization between new and traditional energy sources, achieving a dual improvement in both economic and low-carbon goals. Existing technologies employ a single incentive mechanism for new energy (such as fixed subsidies), making it difficult to balance system economics and low-carbon performance. This solution innovatively designs a "new energy incentive weighting coefficient α" to construct a two-way adjustment mechanism: on the one hand, a negative weighting converts new energy (wind power, photovoltaic) output into system cost deductions, proactively tapping into the potential for new energy consumption; on the other hand, a positive weighting highlights the cost of traditional high-carbon equipment (gas boilers), curbing excessive fossil fuel consumption. This mechanism breaks through the limitations of traditional single-target scheduling, reducing carbon emissions while lowering total operating costs, and promoting the green and low-carbon transformation of virtual power plants—a synergistic effect that existing technologies focusing solely on economic costs or single new energy consumption cannot achieve.

[0241] 3. This solution addresses the technical issues of "single type of flexible load dispatching and insufficient electricity / heat coordination," fully tapping the flexibility potential of the load side. Existing technologies mostly focus on a single type of flexible load (such as only considering loads that can be reduced) and fail to achieve coordinated dispatching of electricity / heat loads, thus failing to fully utilize the load-side regulation capacity. This solution constructs a coordinated model for multiple types of electricity / heat loads that can be shifted, transferred, and reduced, guiding load optimization through differentiated compensation coefficients: shiftable loads enable time-based transfers, transferable loads enable spatiotemporal adjustments, and reduceable loads address emergency scenarios; simultaneously, it establishes power balance constraints and thermal balance constraints, coordinating the electricity / heat output of gas turbines / boilers with energy storage and the electricity / heat storage of electric vehicles (e.g., supplementing electricity through electric vehicle discharge and supplementing heat through gas boiler heating during peak hours). This design reduces the peak-to-valley load difference, increases the renewable energy absorption rate, and achieves coupled optimization of the electricity / heat system, far exceeding the dispatching effects of existing single load types and single energy forms.

[0242] 4. This solution addresses the technical challenge of high interaction costs between virtual power plants and the power grid, optimizing the electricity purchase and sale strategy. Existing technologies fail to fully utilize the "mobile battery" characteristics of electric vehicles and time-of-use pricing differences, resulting in high grid purchase costs and low sales revenue. This solution utilizes the coordinated charging and discharging of electric vehicles and energy storage systems: during off-peak hours (0:00-8:00), electric vehicles and energy storage systems are controlled to charge and store energy; during peak hours (18:00-22:00), both systems discharge to supplement the system's power demand, reducing the need for high-priced electricity purchases; simultaneously, excess renewable energy is fed back to the grid to generate sales revenue. Experimental data shows that after introducing the electric vehicle dispatch model, grid purchase costs are reduced by up to 6.7%, and the discharge revenue from electric vehicles can offset some charging costs, significantly optimizing the economic efficiency of the interaction between virtual power plants and the grid. This cost optimization effect is unattainable by existing dispatch technologies that do not integrate the charging and discharging characteristics of electric vehicles.

[0243] It should be pointed out finally that the above embodiments are only used to illustrate the technical solutions of the present application but not to limit it. Although the present application has been described in detail with reference to the above embodiments, it should be understood by those skilled in the art that the specific embodiments of the present application can be modified or replaced equivalently without departing from the spirit and scope of the present application, and any modification or equivalent replacement without departing from the spirit and scope of the present application should be covered in the protection scope of the claims of the present application.

Claims

1. A virtual power plant optimization scheduling method considering electric vehicle penetration rate and flexible load, characterized in that, Includes the following steps: Step S1: Collect diverse and heterogeneous data required for the operation of the virtual power plant, and perform outlier processing, time scale alignment, and data verification to form a standardized dataset. The diverse and heterogeneous data includes new energy output forecast data, user electricity / heat load forecast data, electric vehicle parameters, and market environment data. Step S2 involves loading equipment cost coefficients and flexible load compensation coefficients based on a standardized dataset, configuring multiple electric vehicle penetration rate scenarios according to the regional level of transportation electrification development, and setting initial values ​​for new energy incentive weight coefficients. Step S2 includes the following steps: Step S21: Based on standardized data, load the operation and maintenance coefficient of new energy units, the fuel coefficient of traditional units, the operation and maintenance coefficient of energy storage systems, and the flexible load compensation coefficient. Step S22: Based on the national transportation electrification planning goals and the current status of regional electric vehicle development, five typical electric vehicle penetration rate levels are set: 0, 0.2, 0.3, 0.4, and 0.

5. Step S23: Set the initial value range of the new energy incentive weight coefficient α to 0.05-0.

2. Incentivize the consumption of new energy through negative weight and suppress the consumption of traditional energy through positive weight. Step S24: Form a structured set of "initialization parameters and configuration scenarios"; Step S3: Based on the initialization parameters and configuration scenario, a virtual power plant optimization scheduling model is constructed with the objective function of minimizing the total operating cost of the virtual power plant, and a complete constraint system is established. The optimal scheduling parameters are obtained through optimization algorithms. The virtual power plant optimization scheduling model comprehensively considers both economic and low-carbon objectives. Step S3 includes the following steps: Step S31: Construct an objective function with the core objective of minimizing the total 24-hour operating cost of the virtual power plant. This function comprehensively covers operation and maintenance costs, fuel costs, electricity purchase costs, electricity sales revenue, energy storage costs, flexible load compensation costs, carbon trading costs, and electric vehicle charging and discharging costs / revenues. In step S31, the introduced new energy incentive weight coefficient α achieves synergistic optimization of economic and low-carbon objectives through the following mechanism: When the system's new energy output is sufficient, increase the negative weight value of the new energy incentive weight coefficient α to convert new energy output into system cost deduction, thus incentivizing the consumption of new energy; when the system's power supply is tight, increase the positive weight value of the new energy incentive weight coefficient α to highlight the cost of traditional high-carbon equipment and suppress excessive consumption of fossil fuels. Step S32: Establish power balance constraints, thermal balance constraints, equipment operation constraints, energy storage system constraints, and electric vehicle charging and discharging constraints to form a complete constraint system; Step S33: Call the Cplex solver through the Yalmip toolbox to solve the objective function under all constraints, obtain the optimal scheduling parameters in hours, and output the optimal power output plan of the equipment and the optimal charging and discharging strategy of electric vehicles for each time period. Step S4: The optimal scheduling parameters obtained by solving are transformed into specific instructions that can be executed by each controllable device, and different charging and discharging strategies are adopted according to different electric vehicle penetration scenarios to issue operation instructions to new energy units, traditional units, energy storage systems and electric vehicle clusters. Step S5: Monitor the actual operating status of each controllable device in real time, compare and analyze the monitoring data with the optimal scheduling parameters, take corresponding adjustment measures according to the degree of deviation, and optimize the parameters based on the actual operating effect after the daily scheduling ends, so as to achieve continuous improvement of the scheduling strategy.

2. The virtual power plant optimization scheduling method considering electric vehicle penetration rate and flexible load according to claim 1, characterized in that, Step S1 includes the following steps: Step S11: Simultaneously collect the new energy output forecast curve and traditional unit parameters on the energy side, the user electricity / heat load forecast curve and flexible load characteristic parameters on the load side, the number of electric vehicles, battery parameters and user travel patterns on the electric vehicle side, and the time-of-use electricity price curve and carbon trading parameters on the market side. Step S12: First, outlier identification and removal are performed on the collected raw data. Then, data with different time resolutions are unified to a 1-hour time step. Finally, data integrity and rationality verification are performed. Step S13: Integrate all the data processed in step S12 into a structured standardized dataset. The standardized dataset includes 1-hour step standardized data from the energy side, load side, electric vehicle side, and market and environment side.

3. The virtual power plant optimization scheduling method considering electric vehicle penetration rate and flexible load according to claim 1, characterized in that, Based on the five typical electric vehicle penetration rate levels of 0, 0.2, 0.3, 0.4, and 0.5 set in step S22, the logic for adjusting the charging and discharging strategy under different electric vehicle penetration rate scenarios is as follows: In a low-penetration scenario with a penetration rate of 0.2, electric vehicles are only charged during off-peak electricity hours from 0:00 to 8:00, with the SOC maintained at 0.5-0.8, and do not participate in discharging during peak hours. In scenarios with low to medium penetration rates (0.3 / 0.4), electric vehicles charge at full power during off-peak electricity hours (0:00-8:00), increasing the State of Charge (SOC) to 0.7-0.

9. During periods of flat electricity prices (8:00-18:00 / 22:00-24:00), charging occurs when the SOC is <0.

6. During peak electricity prices (18:00-22:00), electric vehicles discharge at 20%-30% of their total capacity when the SOC is >0.

8. In a high-penetration scenario with a penetration rate of 0.5, electric vehicles are charged at full load during off-peak hours, with the SOC increasing to 0.8-1.0, and discharged at 30%-50% of the total capacity during the peak electricity price period from 8:00 to 22:

00.

4. The virtual power plant optimization scheduling method considering electric vehicle penetration rate and flexible load as described in claim 1, characterized in that, The objective function is: ， ， ， ， ， ， ， ， in, The cost of operation and maintenance of new energy units, which include wind power (WT) and photovoltaic (PV); These are the WT maintenance coefficient and the PV maintenance coefficient, respectively. These represent the outputs of WT and PV during time period t, respectively. The fuel cost of conventional generating units, which include gas turbines (GT) and gas-fired boilers (GB); These are the GT and GB fuel coefficients, respectively. These represent the outputs of GT and GB during time period t, respectively. For electricity purchase costs; Let t be the power purchased from the grid during time period t. The electricity purchase price for period t; For revenue from electricity sales; Let t be the amount of electricity sold to the grid during time period t. The electricity price for period t; For the operation and maintenance costs of the energy storage system (ESS); This is the ESS operation and maintenance coefficient. Charging power for ESS For ESS discharge power, For ESS thermal storage capacity; For ESS heat dissipation power; To compensate for costs related to flexible loads; For carbon trading costs; Cost of charging electric vehicles; For the discharge revenue of electric vehicles; The charging power of electric vehicles during time period t. The discharge power of the electric vehicle during time period t; For new energy incentive items; This is a fossil fuel constraint.

5. The virtual power plant optimization scheduling method considering electric vehicle penetration rate and flexible load according to claim 4, characterized in that, The power balance constraint is: ， in, Let t be the basic electrical load; The electrical load can be transferred at time t; The electrical load can be shifted at time t; The electrical load can be reduced at time t; The thermodynamic balance constraint is: ， in, Let t be the heating power of the gas turbine. Let t be the heating power of the gas-fired boiler. Let t be the basic thermal load. The thermal load can be shifted at time t. Transfer of thermal load at time t The thermal load can be reduced at time t; The operating constraints of the equipment are: ， in, This refers to the rated power of the gas turbine. For WT to predict maximum output, For GB's predicted maximum output, For PV, the predicted maximum output To minimize the power exchanged with the power grid, This represents the maximum power exchanged with the power grid. The constraints of the energy storage system are: ， in, This represents the minimum charging / discharging power of the energy storage system. This represents the maximum charging / discharging power of the energy storage. This represents the minimum energy storage / heat release power. This represents the maximum energy storage / heat release power. For energy storage charging status, The energy is stored and discharged, and charging and discharging do not occur simultaneously; The charging and discharging constraints for the electric vehicle are: ， in, The maximum charging power for ESS. This represents the maximum discharge power of the ESS.

6. The virtual power plant optimization scheduling method considering electric vehicle penetration rate and flexible load according to claim 1, characterized in that, Step S4 includes the following steps: Step S41: The obtained optimal scheduling parameters in hourly units are converted into control signals that can be recognized by the equipment controller. The control signals include output commands for new energy units, start-stop commands for traditional units, charging and discharging commands for energy storage systems, and charging and discharging strategies for electric vehicle clusters. Step S42: Adopt differentiated charging and discharging strategies according to different electric vehicle penetration scenarios. In low penetration scenarios, priority is given to ensuring users' travel needs, while in high penetration scenarios, the mobile energy storage characteristics of electric vehicles are fully utilized. Step S43: The virtual power plant central control system issues operating instructions to each distributed resource to ensure the effective execution of the scheduling plan.

7. The virtual power plant optimization scheduling method considering electric vehicle penetration rate and flexible load according to any one of claims 1-6, characterized in that, Step S5 includes the following steps: Step S51: Collect the actual operating data of each device in real time through the SCADA system. The actual operating data includes the actual output of new energy, the actual output of traditional units, the actual status of energy storage system and the actual charging and discharging behavior of electric vehicles. Step S52: Compare the actual running data with the obtained optimal scheduling parameters, calculate the degree of deviation, and analyze the reasons for the deviation. Step S53: Based on the deviation analysis results, if the deviation is less than 5%, maintain the original instruction; if the deviation is greater than or equal to 5%, re-execute the model solving step to generate a new instruction. Step S54: After the daily scheduling is completed, update the new energy incentive weight coefficient and electric vehicle charging and discharging strategy based on the actual operation results to achieve continuous optimization of the scheduling model.

Images