Day-ahead optimization scheduling method for virtual power plant containing multiple types of loads

Abstract

The invention discloses a day-ahead optimal scheduling method for a virtual power plant containing multiple types of loads. The day-ahead optimal scheduling method comprises the following steps: step 1, investigating and analyzing the power plant; step 2, collecting load data; 3, constructing a model; step 4, optimizing a scheduling strategy; step 5, plan transmission response; according to the method, the time-of-use electricity price of the user side is associated with different types of users, so that the electricity utilization cost of the users is reduced while the benefits of the virtual power plant are improved; establishing a virtual power plant day-ahead optimization scheduling model containing multiple types of loads and designing time-of-use electricity price pricing strategies of three different types of users to obtain an optimal scheduling strategy based on the time-of-use electricity price; and transmitting the peak-valley flat price data to different types of users for load response, distributing a next-day output plan to each power generation resource for actual power generation, calling an internal power generation side and a user side of the virtual power plant while considering carbon emission permit transaction, realizing optimal configuration of internal resources of the virtual power plant, and promoting low-carbon economic operation of the virtual power plant.

Description

Technical Field

[0001] The present invention relates to the technical field of virtual power plants, and in particular to a day-ahead optimization scheduling method for virtual power plants containing multiple types of loads. Background Art

[0002] Virtual power plant optimization and dispatching is a crucial link in the power system. It involves optimizing the operating parameters of various distributed energy equipment within the virtual power plant to maximize economic benefits while meeting load demand and market electricity prices. At this stage, with the development of dual carbon goals and the construction of new power systems, the proportion of new energy continues to increase, but new energy power generation has the characteristics of randomness, intermittency and volatility. After the new power system introduces large-scale new energy, the balance of power supply and demand and frequency stability will face severe challenges. Traditional power regulation methods can no longer bear this heavy responsibility alone, and there is an urgent need to deeply explore the more efficient and flexible load-side resource response potential.

[0003] In this context, Virtual Power Plant (VPP) has emerged as an innovative and effective solution. It serves as an interactive channel between users and large-scale power grids. It uses the integrated application of "cloud-edge collaboration, Internet of Things technology, and artificial intelligence" to effectively aggregate the flexible response resources on the load side that are widely distributed, numerous, but have small individual capacities, so that resources are observable, measurable, and adjustable at all times, and participate in the electricity and carbon trading markets, which effectively promotes the interaction and collaborative operation of the new power system in multiple links of source-grid-load-storage. The existing virtual power plant optimization scheduling method basically meets the demand, but there are still certain shortcomings: First, the existing virtual power plant scheduling method The study did not link the user-side time-of-use electricity price in the demand-side response with the multiple types of loads within the virtual power plant. To a certain extent, it was unable to effectively tap the user's flexible response potential, which inhibited the development of virtual power plants. Second, the existing dispatch method research failed to provide the optimal dispatch strategy for virtual power plants based on the actual conditions of the peak, valley and normal periods of the electricity market, electricity price information, and elasticity coefficients of various types of loads. Third, the existing dispatch method research failed to simultaneously cover the internal power generation resources of the virtual power plant and the loads guided by time-of-use electricity prices, which affected the low-carbon economic operation of the virtual power plant. Therefore, it is necessary to design a virtual power plant with multiple types of loads. Day-ahead optimization dispatching method. Summary of the invention

[0004] The purpose of the present invention is to provide a method for optimizing the day-ahead scheduling of a virtual power plant containing multiple types of loads, so as to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions: a method for optimizing the day-ahead scheduling of a virtual power plant containing multiple types of loads, comprising the following steps: step 1, power plant investigation and analysis; step 2, load data collection; step 3, model construction; step 4, optimizing the scheduling strategy; step 5, planning the delivery response;

[0006] In the above step 1, based on the centralized spot market, the types of virtual power plants are investigated, which can be divided into demand-side, supply-side and hybrid types according to internal resources;

[0007] In the above step 2, the peak-valley power demand price elasticity coefficients of different types of users of the virtual power plant operator are calculated, and the users within the virtual power plant are classified into Class A industrial users, Class B industrial companies, commercial users and residential users; Class A industrial users refer to industrial users with high sensitivity to electricity prices and large adjustable load potential, and Class B users have lower responsiveness to electricity prices than Class A industrial users. The elasticity coefficient of the users is calculated through the load data before and after the implementation of the peak-valley time-of-use electricity price;

[0008] In the above step 3, peak and valley time-of-use electricity prices are implemented for the internal load of the virtual power plant, the peak and valley electricity price response behavior of users is modeled, and the carbon trading mechanism is modeled;

[0009] In the above step 4, three pricing strategies for time-of-use electricity prices for multiple types of users are proposed. At the same time, the virtual power plant day-ahead optimization scheduling model containing multiple types of loads belongs to a mixed integer nonlinear programming problem. It is solved by optimization software to obtain the optimal scheduling strategy for the virtual power plant day-ahead.

[0010] In the above step five, the model obtains the optimal time-of-use electricity price data for various types of users within the virtual power plant on the next day, transmits the electricity price information to different users through the communication network, and also obtains the internal power generation resource output plan, which is then distributed to each internal member.

[0011] As a further technical solution of the present invention, in the step one, the power plant type is specifically: the demand-side resource type is mainly dominated by the user's own decision-making through transferable and reducible loads based on electricity prices and interruptible loads based on incentives; the supply-side resource type focuses on the peak or valley period of the power grid load, and realizes dynamic regulation of the power supply status by flexibly adjusting the output power of the generator set or optimizing the charging and discharging process of the energy storage equipment to ensure the stable operation of the power grid and the balance of supply and demand; the core of the hybrid resource type is to use advanced energy management systems for intelligent optimization and control, aiming to maximize energy utilization and system economic benefits; the importance of virtual power plants in formulating user peak and valley time-of-use electricity prices and electricity prices is studied; the demand-side response is divided into price type and incentive type, among which the price type aims to directly transmit market price change signals to terminal power users through the implementation of differentiated electricity price policies, such as time-of-use electricity prices, real-time electricity prices and peak electricity prices, and users adjust their electricity usage methods according to changes in electricity prices and adopt peak-shifting Electricity consumption can help reduce or transfer some peak loads, thereby achieving efficient use of energy; price-based demand-side response can, to a certain extent, accurately transmit price signals in the electricity supply and demand relationship to users, promote the operation of virtual power plants in a market-oriented manner, enable virtual power plants to more flexibly dispatch and manage decentralized distributed energy resources, and promote efficient use of resources; time-of-use electricity prices belong to the category of price-based demand response, which can guide users to shave peaks and fill valleys through differences in electricity prices in different periods, use electricity scientifically and reduce costs at the same time; the peak and valley time-of-use electricity prices on the user side of virtual power plants should be divided into time periods and the price mechanism should be designed with key considerations on the potential for regulating electricity load characteristics, the need to connect with the spot market, and the degree of acceptability of market players. Time-of-use electricity prices, with their fixed electricity prices within time periods, provide users with clear and predictable price signals. This feature does not require users' metering equipment and response awareness, and has high acceptability and response convenience at the user level, which helps to tap users' response potential and promote efficient use of resources.

[0012] As a further technical solution of the present invention, in step 2, specifically, the model considers that the user's electricity consumption in a certain period of time will be affected by the comprehensive influence of the electricity price in the current period and the electricity price in other periods, so the peak-valley electricity demand elasticity coefficient expression is:

[0013]

[0014] where ε ii is the elastic coefficient, This section shows the electricity price and electricity quantity. Δδ i Respectively represent the amount of change;

[0015] The electricity price elasticity coefficient ε ij There are two types: self-elasticity and mutual elasticity:

[0016]

[0017] In the formula, when j = i, ε ij is the elastic coefficient, j≠i, ε ij is the mutual elastic coefficient; δ j They represent the electricity quantity and electricity price in the corresponding time period respectively. Δδ j They represent the changes in the corresponding time periods respectively;

[0018] The response behavior of users to electricity prices in period n is modeled as follows:

[0019]

[0020] Among them, n is the number of time periods, the diagonal elements are the self-elasticity coefficients, and the non-diagonal elements are the cross-elasticity coefficients; various types of load data are collected in real time through intelligent monitoring equipment. Based on this, according to industry characteristics and electricity consumption patterns, users are accurately divided into four categories: Class A industrial users, Class B industrial users, commercial users and residential users; virtual power plant operators use advanced prediction algorithms and big data analysis capabilities to fully analyze the load characteristics of various types of users, such as peak values, valley values, peak-to-valley differences, and load rates, study the changing laws of load curves, and analyze various factors that affect load changes, such as meteorological conditions (temperature, humidity, wind speed, etc.), holidays, special events (power outages, maintenance, etc.), socio-economic factors, etc., and obtain typical daily load curves of various types of loads through the K-means clustering algorithm, and merge them into the total load curve of the virtual power plant operator.

[0021] As a further technical solution of the present invention, in the step 2, the peak, valley and average sections of the load curve are divided. Specifically, the peak, valley and average sections of the load are divided by deeply analyzing the peak and valley characteristics of various load curves, and defining the extreme value intervals of each load curve, that is, the highest load point and the lowest load point, and then using the fuzzy boundary gradient membership function mechanism to scientifically divide the peak, valley and average sections of the load; assuming that the set of time points on the typical daily load curve of the user is T = {i1 i2 … i 24}, the value set corresponding to each period on the curve is L = l1 l2 … l 24 , then the valley time point set corresponding to the valley period is T v , the set of ordinary points corresponding to the ordinary period is T f , the peak time point set corresponding to the peak period is T p , then the valley membership at each time point can be calculated by the fuzzy semi-gradient membership function Sum peak membership Right now;

[0022]

[0023] i∈T v

[0024]

[0025] i∈T p

[0026] In order to more accurately calculate the degree of membership at each time point, the present invention replaces l in the formula i It is extended to mean the average of the loads corresponding to time i and time i+1.

[0027] As a further technical solution of the present invention, in step 3, the behavior modeling is specifically that when the electricity price is high during a certain period of time, the user actively responds by reducing the current unnecessary power load or transferring part of the power demand to a period with a more favorable electricity price. Therefore, in the load response mode of 24 periods a day, the load transfer and reduction variables are:

[0028]

[0029] Where T p , T f , T v are the peak, flat and valley periods of load, respectively, pf , pv , fv are the interaction transfer coefficients of each period, λ pp , ff , vv are the reduction coefficients for each period, δ p ,δ f ,δ v are the prices during peak, flat and valley periods respectively, k p , k f , k v are the electricity price fluctuation ratios in the corresponding time periods, and k p >0, k f =0, k v <0, so λ ff =0,λ fv =0,λ pf >0,λ pv >0, in addition, due to ε ii < 0, λ pp >0,λ vv <0, we get:

[0030]

[0031] Where L' t is the load in period t after execution, L t is the load of the original period t, L p , L f , L vare the average load values ​​for each period.

[0032] As a further technical solution of the present invention, in the step 3, the VPP components proposed by the present invention include distributed wind power, distributed photovoltaics, micro gas turbines, distributed energy storage and adjustable loads; among which distributed photovoltaics and wind power belong to clean renewable energy, so the carbon emission sources mentioned in the present invention include micro gas turbines and power purchased by the distribution network (assuming that they are all generated by thermal power units). At the same time, according to the national mechanism for voluntary emission reductions, in addition, according to the national mechanism for voluntary emission reductions, the environmental protection contributions of wind power and photovoltaics are included in the calculation scope of the free carbon emission quota; the free carbon emission amount of VPP is:

[0033]

[0034] Where α is the regional carbon emission quota per unit of electricity, is the power generation of the gas turbine at time t, For wind power, Produce electricity for photovoltaics;

[0035] The actual carbon emissions of VPP are:

[0036]

[0037] Among them, β1 is the carbon emission coefficient of the micro gas turbine, β2 is the carbon emission coefficient of the power purchased by the distribution network, is the power purchased by VPP in the distribution network at time t;

[0038] The present invention sets up VPP to allow participation in carbon emission rights trading. Under this mode, VPP can flexibly participate in the purchase and sale of carbon emission rights in the carbon trading market, thereby optimizing the configuration and management of carbon assets according to its own operating strategy and market conditions. The carbon trading cost is:

[0039]

[0040] where δ c It is the carbon trading price of the day.

[0041] As a further technical solution of the present invention, in step 3, the specific process of optimizing the scheduling model is as follows: the characteristics of the internal power generation resources of the virtual power plant, the user's electricity cost for:

[0042]

[0043] in is the electricity price sold by VPP to user n during period t, is the initial power consumption of user n in period t, is the change in electricity consumption of user n in time period t after responding to the electricity price, and N is the number of user types;

[0044] Grid transaction costs for:

[0045]

[0046] in is the amount of electricity purchased from the distribution network at time t, is the electricity price corresponding to the electricity purchased from the distribution network at time t, is the amount of electricity sold to the distribution network at time t, is the electricity price corresponding to the electricity sold to the distribution network at time t;

[0047] Gas Turbine Cost for:

[0048]

[0049] in is the power generation of the gas turbine at time t, and a, b, and c are their cost coefficients respectively;

[0050] Energy storage compensation costs for:

[0051]

[0052] where δ ev is the storage unit compensation price, is the charging and discharging power of energy storage participating in the electricity market during period t;

[0053] Wind power penalty costs Photovoltaic penalty costs for:

[0054]

[0055] in is the deviation between the actual wind power output and the predicted output during period t, is the penalty price of wind power in period t, is the deviation between the actual photovoltaic output and the predicted output during period t, is the penalty price of photovoltaic power generation during period t;

[0056] Time-of-use electricity price compensation for users for:

[0057]

[0058] in is the load reduction transfer amount of the user in the period n to t, is the compensation price corresponding to user n in period t.

[0059] As a further technical solution of the present invention, in step 3, in the day-ahead market transaction, the VPP revenue includes the revenue from selling electric energy in the day-ahead market and selling electric energy to the VPP load, and the VPP cost includes the day-ahead power market purchase cost, gas turbine power generation cost, energy storage compensation cost, wind power and photovoltaic penalty cost, carbon trading cost, and user transfer and load reduction compensation cost; the VPP revenue R is:

[0060]

[0061] The maximum value of VPP revenue maxR of the objective function of the day-ahead optimization dispatch model of a virtual power plant with multiple types of loads is:

[0062]

[0063] in is the revenue of VPP selling electric energy to various types of loads at time t, i.e., the load electricity cost, is the transaction cost of VPP in the day-ahead power market and distribution network at time t, is the gas turbine power generation cost at time t, is the energy storage compensation cost at time t, is the penalty cost of wind power at time t, is the photovoltaic penalty cost at time t, is the carbon trading cost, The time-of-use electricity price compensation cost for users, T is the 24-hour period of the day-ahead dispatch cycle;

[0064] The constraints of the virtual power plant day-ahead optimal dispatch model with multiple types of loads are constructed. The constraints of the model are power balance constraints, gas turbine constraints, wind and solar output constraints, energy storage constraints, electricity market constraints and user satisfaction constraints. The power balance constraint of the virtual power plant at time t is:

[0065]

[0066] in is the amount of electricity purchased from the distribution network at time t, is the amount of electricity sold to the distribution network at time t, is the power generation of the gas turbine at time t, is the charging and discharging power of energy storage participating in the electric energy market during period t, For photovoltaic power generation, To contribute to wind power, is the initial power consumption of user n in period t, is the load reduction transfer amount of user n in period t;

[0067] The gas turbine constraints are:

[0068]

[0069] in is the minimum output of the gas turbine, is the maximum output of the gas turbine, is the maximum downward ramp rate of the gas turbine, is the maximum upward climbing rate of the gas turbine;

[0070] Wind power and photovoltaic output constraints are:

[0071]

[0072] in is the predicted wind power output in period t, δ1 is the wind power prediction error, is the photovoltaic predicted output during period t, η2 is the photovoltaic prediction error;

[0073] The energy storage constraint is:

[0074]

[0075] Where P es,max is the maximum charging and discharging power of energy storage, is the charging and discharging power of energy storage participating in the electric energy market during period t; when the energy storage device participates in the VPP operation optimization, it must meet the requirement that only charging or discharging is allowed during a certain period of time. The specific constraints are:

[0076]

[0077] in is the energy storage charging power during period t, is the energy storage discharge power during period t, U t It is a 0-1 variable, which means that it is 1 if the current period is charged and 0 if it is not charged;

[0078]

[0079] in is the initial charge of the energy storage device, is the final state charge of the energy storage device, is the amount of electricity in the energy storage device during period t, E es,min is the minimum capacity of the energy storage device, E es,max is the maximum capacity of the energy storage device, η c is the charging efficiency of the energy storage device, η d is the discharge efficiency of the energy storage device;

[0080] The electricity market constraints are:

[0081]

[0082] in is the maximum power purchase, is the maximum power sold, is the amount of electricity purchased from the distribution network at time t, is the corresponding amount of electricity sold to the distribution network at time t, B t It is a 0-1 variable, indicating the VPP's participation in the electricity market, 1 represents electricity purchase, and 0 represents electricity sales;

[0083] The user satisfaction constraint is:

[0084]

[0085] refers to the electricity satisfaction of user n, It is the lower limit of electricity satisfaction, and its value is 0.8.

[0086] As a further technical solution of the present invention, in step 4, the specific process is to propose three pricing strategies for different types of users to verify the effectiveness of the virtual power plant day-ahead optimization scheduling model;

[0087] Strategy 1: All users do not use peak and valley time-of-use electricity prices, that is, use flat electricity prices;

[0088] Strategy 2: Only industrial users adopt peak and valley time-of-use electricity prices;

[0089] Strategy 3: Industrial, commercial and residential users all adopt peak and valley time-of-use electricity prices according to their own response capabilities;

[0090] First, Strategy 1 is the strategy currently adopted next to the virtual power plant, which is used as a reference;

[0091] Secondly, Strategy 2 is a transition strategy between Strategy 1 and Strategy 3. It can be proved that the profit of virtual power plant under Strategy 2 is greater than that under Strategy 1.

[0092] Finally, Strategy 3 is designed based on the differences in the response capabilities of different types of users to electricity prices. It can fully tap the flexible response potential of different types of users. At the same time, the benefits of the virtual power plant are greater than those of Strategy 2.

[0093] According to the user load demand, the internal power generation resource parameters (including wind, solar, micro gas turbines, energy storage), time-of-use electricity prices and peak and valley periods in the electricity energy market, and the elasticity coefficients of various types of users are input into the model as input data for the day-ahead optimization scheduling model of the virtual power plant containing multiple types of loads. The model is solved using optimization software to perform optimization and solution. The virtual power plant operator's revenue, user revenue, real-time market revenue of the power sales company, peak electricity prices when different types of users implement time-of-use electricity prices, and the degree of user's active acceptance are obtained under different strategies, and finally the optimal strategy for the day-ahead scheduling of the virtual power plant is determined.

[0094] As a further technical solution of the present invention, in the step five, the specific process is to transmit the peak, valley and flat electricity price data to different types of users for load response. First, the prerequisite is prepared; the user needs to have intelligent devices such as smart meters, load control terminals, data interaction and communication equipment; second, the virtual power plant allocates output plans to power generation resources; the virtual power plant uses advanced information and communication technologies, intelligent control systems, and remote monitoring and scheduling equipment to achieve data exchange and command transmission with power generation resources. Through these technologies, the virtual power plant can send the output plan to the power generation resources in the form of instructions, and collect the operating status, output capacity and other information of the power generation resources in real time, and promptly correct and adjust the behavior that deviates from the plan; third, the power generation resources plan output; after receiving the output plan instruction sent by the virtual power plant, the power generation resources will adjust their own operating status and output level according to the requirements in the instruction. In the process of executing the planned output, the power generation resources will monitor their own operating status and output in real time, and feedback the relevant information to the virtual power plant. The virtual power plant makes necessary adjustments and optimizations to the output plan based on the feedback information to ensure the stable operation and supply and demand balance of the entire power system.

[0095] Compared with the prior art, the beneficial effects of the present invention are as follows: the day-ahead optimization scheduling method for a virtual power plant with multiple types of loads can effectively and reasonably improve the revenue and market competitiveness of the virtual power plant and reduce the risk of the virtual power plant participating in the market based on the actual power market environment; the user-side time-of-use electricity price is linked to different types of users to a certain extent, which helps to improve the revenue of the virtual power plant while reducing the user's electricity cost indirectly, and the user's active acceptance is strong; while establishing a day-ahead optimization scheduling model for a virtual power plant with multiple types of loads, a pricing strategy for time-of-use electricity prices for three different types of users is also designed, and the two are mutually embedded to obtain the optimal scheduling strategy suitable for the virtual power plant and internal load based on time-of-use electricity prices; after obtaining the day-ahead optimal scheduling strategy for a virtual power plant with multiple types of loads, the peak, valley and flat electricity price data are transmitted to different types of users for load response, and the next day's power generation plan is allocated to each power generation resource for actual power generation. While considering carbon emission rights trading, the internal power generation side and user side of the virtual power plant are mobilized to achieve optimal allocation of internal resources of the virtual power plant and promote the low-carbon economic operation of the virtual power plant. BRIEF DESCRIPTION OF THE DRAWINGS

[0096] Figure 1 is an overall flow chart of an embodiment of the present invention;

[0097] Figure 2 This is a schematic diagram of the electricity selling and purchasing prices during a time period in an embodiment of the present invention;

[0098] Figure 3 Schematic diagram of wind power and photovoltaic power prediction output in an embodiment of the present invention;

[0099] Figure 4 Schematic diagram of actual power output of wind power and photovoltaic in the embodiment of the present invention;

[0100] Figure 5 Schematic diagram of total hourly load of Strategy 1, Strategy 2 and Strategy 3 in the embodiment of the present invention;

[0101] Figure 6 Schematic diagram of data such as energy storage charge and discharge in the embodiment of the present invention;

[0102] Figure 7 Schematic diagram of user classification and corresponding elasticity coefficients in the embodiment of the present invention;

[0103] Figure 8 Schematic diagram of parameters of gas turbines and energy storage devices in the embodiment of the present invention;

[0104] Fig. 9 Schematic diagram of time-of-use electricity prices of Strategy 2 and Strategy 3 in the embodiment of the present invention;

[0105] Fig.10 Schematic diagram of benefits and costs under each strategy in the embodiment of the present invention;

[0106] Fig.11 Flowchart of the method of the present invention. Detailed implementation manners

[0107] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0108] Please refer to the attached Figure 1 - attached Fig.10 , an embodiment provided by the present invention: A day-ahead optimal scheduling method for a virtual power plant with multiple types of loads: A detailed description of the day-ahead optimal scheduling method for a virtual power plant with multiple types of loads disclosed;

[0109] First, assume the test conditions;

[0110] Assume that the user types of the VPP are as Figure 8 shown. 1. In terms of load parameters; the sensitivity of the load to time-of-use electricity prices: residential load < commercial load < Class B industrial load < Class A industrial load; 2. In terms of electricity price information: the corresponding prices for the VPP to buy and sell electricity in the electricity energy market are as Figure 2As shown; wherein, the peak and valley time periods of the distribution network are divided into: the peak period is 11:00-18:00, the flat period is 7:00-10:00, 19:00-22:00, and the valley period is 0:00-6:00, 23:00-24:00; 3. According to the relevant provisions of the "Implementation Rules for Power Forecast and Forecast Management of Wind Farms", the present invention sets the power forecast error rates of wind farms and photovoltaic power stations to be 12.05% and 8.02% respectively; the wind and solar output forecast curves and the typical daily load curves of different types of users are as follows Figure 3 As shown in Figure 2, the wind power and photovoltaic output curves are as follows: Figure 4 4. Parameters related to the virtual power plant micro gas turbine and energy storage equipment are as follows: Figure 8 As shown; the three proposed pricing strategies are optimized and simulated, and the results are as follows Figure 5 , Fig. 9 and Fig.10 As shown; 5. The impact of each strategy on the peak-shaving and valley-filling curve; Figure 5 and Fig. 9 As shown in the figure: Compared with Strategy 1 and Strategy 2, in Strategy 2, only industrial users adopt time-of-use electricity prices, and the industrial user satisfaction is higher when the model obtains the optimal time-of-use electricity price, but it fails to fully release the potential of industrial users to participate in demand response, so it is relatively limited in adjusting the power load curve to achieve peak shaving and valley filling; Compared with Strategy 2 and Strategy 3, Strategy 3 is that the time-of-use electricity price covers all user groups. This comprehensiveness not only significantly enhances the enthusiasm of industrial users to participate in demand response and deeply taps their potential, but also significantly reduces the load peak-valley difference and increases the load rate through a wider user base. Its effect is significantly better than Strategy 2; 6. Benefit comparison; The benefits and costs of VPP under the three strategies are as follows Fig.10 As shown in the figure, only industrial users participate in strategy 2, and the virtual power plant obtains limited benefits; while strategy 3 is to take into account user satisfaction while all users adopt time-of-use electricity prices. Users increase their electricity costs, but they have been offset by virtual power plant compensation while increasing profits; the virtual power plant aggregates the total load and power generation resources after the user's time-of-use electricity price and trades in the day-ahead market to earn a certain amount of income, so as to minimize the transaction costs of the distribution network and the carbon market; finally, on the basis of ensuring the original income of VPP in strategy 1 and compensating for the increase in user electricity costs, compared with strategy 1, an additional 1,046.9 yuan of income is added, and at the same time, the user earns 287 yuan through virtual power plant compensation without changing the cost of purchasing electricity; 7. Analysis of the output of each VPP entity under the optimal strategy; in strategy 3, the virtual power plant operator analyzes and makes decisions on the basic information of the day-ahead electricity energy market, optimizes the scheduling according to the aggregated wind power, photovoltaic, energy storage and other resources, and obtains the output of each resource within the VPP, as shown in the figure. Figure 6 As shown, the analysis shows that:

[0111] 1:00 to 6:00: During this period, photovoltaic power generation is relatively small due to the influence of the external environment. It is difficult to achieve a balance between supply and demand with the load by relying solely on the internal power generation resources of the virtual power plant. The electricity price in the energy market is relatively low during this period. Therefore, VPP mainly acts as a power purchaser and purchases electricity from the distribution network to meet basic load demand.

[0112] Under the influence of time-of-use electricity prices, the charging and discharging periods of energy storage devices also have certain rules. Energy storage charging is mainly carried out during the two periods of 1-6 o'clock and 23-24 o'clock. Under the guidance of price-based demand response, the transferable and reducible loads within the virtual power plant are more likely to choose to use electricity during the electricity price valley period, and generate electricity during normal and peak periods.

[0113] The time period from 11:00 to 18:00 is the peak load period, electricity is in short supply, and the market electricity price reaches the highest. At this time, the photovoltaic power generation inside the virtual power plant is relatively large. At the same time, in order to maximize economic benefits, the virtual power plant will also use energy storage equipment to discharge during the peak power consumption period, guiding users to meet their own electricity satisfaction while transferring and reducing the load during this period, so that the daily load curve during the peak period shows a downward trend, and the valley period shows an upward trend, and the load rate is effectively improved. Therefore, during the peak period, VPP mainly participates in the electricity market as a power seller. Due to the small capacity of distributed resources, the impact on the electricity market price is very small. Therefore, the virtual power plant mainly sells the aggregated power generation energy to the distribution network as a price acceptor in the electricity market, reducing the purchase of electricity from the distribution network during the peak electricity price period and increasing the income from the electricity market;

[0114] Through the above comparison of peaks and valleys of load curves and comparison of benefits, it can be seen that, considering the elasticity coefficient and acceptability of the user-side subject, a multi-type load virtual power plant day-ahead optimization scheduling model with time-of-use electricity price and carbon trading mechanism is established, and three strategies are proposed and case studies are carried out to verify the effectiveness of the model; the results show that in strategy 3, when all users within the virtual power plant operator implement time-of-use electricity price, it can effectively tap the response capabilities of various types of users, improve the benefits of virtual power plants, promote their low-carbon economic operation, and at the same time, smooth the peaks and fill the valleys of the load curve and reduce the user's electricity purchase costs.

[0115] Please see attached Fig.11 , an embodiment provided by the present invention: a method for optimizing the scheduling of a virtual power plant with multiple types of loads, comprising the following steps: step one, power plant investigation and analysis, step two, load data collection; step three, model construction; step four, optimizing the scheduling strategy; step five, planning the delivery response;

[0116] In the above step 1, based on the centralized spot market, the types of virtual power plants are investigated, which can be divided into demand-side, supply-side and hybrid types according to internal resources; the specific types of power plants are as follows: the demand-side resource type is mainly dominated by the user's own decision-making through transferable and reducible loads based on electricity prices and interruptible loads based on incentives; the supply-side resource type focuses on the dynamic regulation of power supply conditions during peak or valley periods of grid load by flexibly adjusting the output power of generator sets or optimizing the charging and discharging process of energy storage equipment to ensure the stable operation of the grid and the balance of supply and demand; the core of the hybrid resource type is to use advanced energy management systems for intelligent optimization and control, aiming to maximize energy utilization and system economic benefits; the importance of virtual power plants in formulating user peak and valley time-of-use electricity prices and dividing time periods and electricity prices is studied; the demand-side response is divided into price type and incentive type, among which the price type aims to directly transmit market price change signals to end-users through the implementation of differentiated electricity price policies, such as time-of-use electricity prices, real-time electricity prices and peak electricity prices, and users can make decisions based on electricity prices. Adjusting electricity consumption patterns according to price changes and adopting staggered electricity consumption can help reduce or shift some peak loads, thereby achieving efficient use of energy; price-based demand-side response can, to a certain extent, accurately transmit price signals in the electricity supply and demand relationship to users, promote virtual power plants to operate in a market-oriented manner, enable virtual power plants to more flexibly dispatch and manage decentralized distributed energy resources, and promote efficient use of resources; time-of-use electricity prices belong to the category of price-based demand response, which can guide users to reduce peak loads and fill valleys through price differences in different periods, use electricity scientifically and reduce costs at the same time; the peak and valley time-of-use electricity prices on the user side of virtual power plants should be divided into time periods and the price mechanism should be designed with key considerations on the potential for regulating electricity load characteristics, the need to connect with the spot market, and the degree of acceptability of market players. Time-of-use electricity prices, with their fixed price characteristics within the time period, provide users with clear and predictable price signals. This feature does not require high metering equipment and response awareness from users, and has high acceptability and response convenience at the user level, which helps to tap the user's response potential and promote efficient use of resources;

[0117] In the above step 2, the peak-valley power demand price elasticity coefficients of different types of users of the virtual power plant operator are calculated, and the users within the virtual power plant are classified into Class A industrial users, Class B industrial companies, commercial users and residential users; Class A industrial users refer to industrial users with high sensitivity to electricity prices and large adjustable load potential, and Class B users have lower responsiveness to electricity prices than Class A industrial users. The elasticity coefficient of the user is calculated by the load data before and after the implementation of the peak-valley time-of-use electricity price; specifically, the model considers that the user's electricity consumption in a certain period of time will be affected by the comprehensive influence of the current period electricity price and other period electricity prices, so the expression of the peak-valley power demand elasticity coefficient is:

[0118]

[0119] where ε iiis the elastic coefficient, This section shows the electricity price and electricity quantity. Δδ i Respectively represent the amount of change;

[0120] The electricity price elasticity coefficient ε ij There are two types: self-elasticity and mutual elasticity:

[0121]

[0122] In the formula, when j = i, ε ij is the elastic coefficient, j≠i, ε ij is the mutual elastic coefficient; δ j They represent the electricity quantity and electricity price in the corresponding time period respectively. Δδ j They represent the changes in the corresponding time periods respectively;

[0123] The response behavior of users to electricity prices in period n is modeled as follows:

[0124]

[0125] Where n is the number of time periods, the diagonal elements are the self-elasticity coefficients, and the non-diagonal elements are the cross-elasticity coefficients; various types of load data are collected in real time through intelligent monitoring equipment. Based on this, users are accurately divided into four categories: Class A industrial users, Class B industrial users, commercial users, and residential users according to industry characteristics and electricity consumption patterns; virtual power plant operators use advanced prediction algorithms and big data analysis capabilities to fully analyze the load characteristics of various types of users, such as peak values, valley values, peak-to-valley differences, and load rates, study the changing laws of load curves, and analyze various factors that affect load changes, such as meteorological conditions (temperature, humidity, wind speed, etc.), holidays, special events (power outages, maintenance, etc.), socio-economic factors, etc., and obtain typical daily load curves of various types of loads through the K-means clustering algorithm, and merge them into the total load curve of the virtual power plant operator;

[0126] The peak, valley and average periods of the load curve are divided. Specifically, the peak, valley and average period division method of the load is to deeply analyze the peak and valley characteristics of various load curves, and define the extreme value interval of each load curve, that is, the highest load point and the lowest load point, and then use the fuzzy boundary gradient membership function mechanism to scientifically divide the peak, valley and average periods of the load; assuming that the set of time points on the typical daily load curve of the user is T = {i1 i2 … i 24}, the value set corresponding to each period on the curve is L = l1 l2 … l 24 , then the valley time point set corresponding to the valley period is T v , the set of ordinary points corresponding to the ordinary period is T f , the peak time point set corresponding to the peak period is Tp , then the valley membership at each time point can be calculated by the fuzzy semi-gradient membership function Sum peak membership Right now;

[0127]

[0128] i∈T v

[0129]

[0130] i∈T p

[0131] In order to more accurately calculate the degree of membership at each time point, the present invention replaces l in the formula i It is extended to mean the mean of the load corresponding to time i and time i+1;

[0132] In the above step 3, the peak-valley time-of-use electricity price is implemented for the internal load of the virtual power plant, the peak-valley electricity price response behavior of the user is modeled, and the carbon trading mechanism is modeled; the VPP components proposed in the present invention include distributed wind power, distributed photovoltaics, micro gas turbines, distributed energy storage and adjustable loads; among which distributed photovoltaics and wind power belong to clean renewable energy, so the carbon emission sources mentioned in the present invention include micro gas turbines and distribution network purchased electricity (assuming that they are all thermal power generation units). At the same time, according to the national certified voluntary emission reduction mechanism, in addition, according to the national certified voluntary emission reduction mechanism, the environmental protection contribution of wind power and photovoltaics is included in the calculation scope of the free carbon emission quota; the free carbon emission amount of VPP is:

[0133]

[0134] Where α is the regional carbon emission quota per unit of electricity, is the power generation of the gas turbine at time t, For wind power, Produce electricity for photovoltaics;

[0135] The actual carbon emissions of VPP are:

[0136]

[0137] Among them, β1 is the carbon emission coefficient of the micro gas turbine, β2 is the carbon emission coefficient of the power purchased by the distribution network, is the power purchased by VPP in the distribution network at time t;

[0138] The present invention sets up VPP to allow participation in carbon emission rights trading. Under this mode, VPP can flexibly participate in the purchase and sale of carbon emission rights in the carbon trading market, thereby optimizing the configuration and management of carbon assets according to its own operating strategy and market conditions. The carbon trading cost is:

[0139]

[0140] where δ c is the carbon trading price of the day;

[0141] The specific process of the optimization scheduling model is as follows: the characteristics of the internal power generation resources of the virtual power plant, the user's electricity cost for:

[0142]

[0143] in is the electricity price sold by VPP to user n during period t, is the initial power consumption of user n in period t, is the change in electricity consumption of user n in time period t after responding to the electricity price, and N is the number of user types;

[0144] Grid transaction costs for:

[0145]

[0146] in is the amount of electricity purchased from the distribution network at time t, is the electricity price corresponding to the electricity purchased from the distribution network at time t, is the amount of electricity sold to the distribution network at time t, is the electricity price corresponding to the electricity sold to the distribution network at time t;

[0147] Gas Turbine Cost for:

[0148]

[0149] in is the power generation of the gas turbine at time t, and a, b, and c are their cost coefficients respectively;

[0150] Energy storage compensation costs for:

[0151]

[0152] where δ ev is the storage unit compensation price, is the charging and discharging power of energy storage participating in the electricity market during period t;

[0153] Wind power penalty costs Photovoltaic penalty costs for:

[0154]

[0155] in is the deviation between the actual wind power output and the predicted output during period t, is the penalty price of wind power in period t, is the deviation between the actual photovoltaic output and the predicted output during period t, is the penalty price of photovoltaic power generation during period t;

[0156] Time-of-use electricity price compensation for users for:

[0157]

[0158] in is the load reduction transfer amount of the user in the period n to t, is the compensation price corresponding to user n in time period t;

[0159] In the day-ahead market transaction, VPP revenue includes the revenue from selling electricity in the day-ahead market and selling electricity to VPP loads. VPP costs include the day-ahead electricity market purchase cost, gas turbine power generation cost, energy storage compensation cost, wind power and photovoltaic penalty cost, carbon trading cost, and user transfer and load reduction compensation cost. VPP revenue R is:

[0160]

[0161] The maximum value of VPP revenue maxR of the objective function of the day-ahead optimization dispatch model of a virtual power plant with multiple types of loads is:

[0162]

[0163] in is the revenue of VPP selling electric energy to various types of loads at time t, i.e., the load electricity cost, is the transaction cost of VPP in the day-ahead power market and distribution network at time t, is the gas turbine power generation cost at time t, is the energy storage compensation cost at time t, is the penalty cost of wind power at time t, is the photovoltaic penalty cost at time t, is the carbon trading cost, The time-of-use electricity price compensation cost for users, T is the 24-hour period of the day-ahead dispatch cycle;

[0164] The constraints of the virtual power plant day-ahead optimal dispatch model with multiple types of loads are constructed. The constraints of the model are power balance constraints, gas turbine constraints, wind and solar output constraints, energy storage constraints, electricity market constraints and user satisfaction constraints. The power balance constraint of the virtual power plant at time t is:

[0165]

[0166] in is the amount of electricity purchased from the distribution network at time t, is the amount of electricity sold to the distribution network at time t, is the power generation of the gas turbine at time t, is the charging and discharging power of energy storage participating in the electric energy market during period t, For photovoltaic power generation, To contribute to wind power, is the initial power consumption of user n in period t, is the load reduction transfer amount of user n in period t;

[0167] The gas turbine constraints are:

[0168]

[0169] in is the minimum output of the gas turbine, is the maximum output of the gas turbine, is the maximum downward ramp rate of the gas turbine, is the maximum upward climbing rate of the gas turbine;

[0170] Wind power and photovoltaic output constraints are:

[0171]

[0172] in is the predicted wind power output in period t, η1 is the wind power prediction error, is the photovoltaic predicted output during period t, η2 is the photovoltaic prediction error;

[0173] The energy storage constraint is:

[0174]

[0175] Where P es,max is the maximum charging and discharging power of energy storage, is the charging and discharging power of energy storage participating in the electric energy market during period t; when the energy storage device participates in the VPP operation optimization, it must meet the requirement that only charging or discharging is allowed during a certain period of time. The specific constraints are:

[0176]

[0177]

[0178] in is the energy storage charging power during period t, is the energy storage discharge power during period t, U t It is a 0-1 variable, which means that it is 1 if the current period is charged and 0 if it is not charged;

[0179]

[0180] in is the initial charge of the energy storage device, is the final state charge of the energy storage device, is the amount of electricity in the energy storage device during period t, E es,min is the minimum capacity of the energy storage device, E es,max is the maximum capacity of the energy storage device, η c is the charging efficiency of the energy storage device, η d is the discharge efficiency of the energy storage device;

[0181] The electricity market constraints are:

[0182]

[0183] in is the maximum power purchase, is the maximum power sold, is the amount of electricity purchased from the distribution network at time t, is the corresponding amount of electricity sold to the distribution network at time t, B t It is a 0-1 variable, indicating the VPP's participation in the electricity market, 1 represents electricity purchase, and 0 represents electricity sales;

[0184] The user satisfaction constraint is:

[0185]

[0186] refers to the electricity satisfaction of user n, It is the lower limit of electricity satisfaction, and its value is 0.8.

[0187] In the above step 4, three pricing strategies for time-of-use electricity prices for multiple types of users are proposed. At the same time, the virtual power plant day-ahead optimization scheduling model containing multiple types of loads belongs to a mixed integer nonlinear programming problem. It is solved by optimization software to obtain the optimal scheduling strategy for the virtual power plant day-ahead. The specific process is to propose three pricing strategies for different types of users to verify the effectiveness of the virtual power plant day-ahead optimization scheduling model.

[0188] Strategy 1: All users do not use peak and valley time-of-use electricity prices, that is, use flat electricity prices;

[0189] Strategy 2: Only industrial users adopt peak and valley time-of-use electricity prices;

[0190] Strategy 3: Industrial, commercial and residential users all adopt peak and valley time-of-use electricity prices according to their own response capabilities;

[0191] First, Strategy 1 is the strategy currently adopted next to the virtual power plant, which is used as a reference;

[0192] Secondly, Strategy 2 is a transition strategy between Strategy 1 and Strategy 3. It can be proved that the profit of virtual power plant under Strategy 2 is greater than that under Strategy 1.

[0193] Finally, Strategy 3 is designed based on the differences in the response capabilities of different types of users to electricity prices. It can fully tap the flexible response potential of different types of users. At the same time, the benefits of the virtual power plant are greater than those of Strategy 2.

[0194] According to the user load demand, the internal power generation resource parameters (including wind and solar, micro gas turbines, energy storage), the time-of-use electricity price and peak and valley time periods in the electricity energy market, and the elasticity coefficients of various types of users are input into the model as the input data of the virtual power plant day-ahead optimization scheduling model containing multiple types of loads. The optimization software is used to solve the model and perform optimization and solution to obtain the virtual power plant operator's revenue, user revenue, real-time market revenue of the power sales company, peak electricity prices when different types of users implement time-of-use electricity prices, and the degree of active acceptance of users under different strategies, and finally determine the optimal strategy for the day-ahead scheduling of the virtual power plant;

[0195] In the above step five, the model obtains the optimal time-of-use electricity price data for various types of users in the virtual power plant on the next day, transmits electricity price information to different users through the communication network, and also obtains the internal power generation resource output plan, and then distributes the output plan to each internal member; the specific process is to transmit the peak, valley and flat electricity price data to different types of users for load response. First, the prerequisite is to prepare; users need to have smart devices such as smart meters, load control terminals, data interaction and communication equipment; second, the virtual power plant allocates output plans to power generation resources; the virtual power plant uses advanced information and communication technology, intelligent control systems, remote monitoring and dispatching equipment to realize data exchange and command transmission with power generation resources, Through these technologies, the virtual power plant can send the output plan to the power generation resources in the form of instructions, and collect information such as the operating status and output capacity of the power generation resources in real time, and promptly correct and adjust the behavior that deviates from the plan; third, the power generation resources plan the output; after receiving the output plan instructions sent by the virtual power plant, the power generation resources will adjust their own operating status and output level according to the requirements of the instructions. In the process of executing the planned output, the power generation resources will monitor their own operating status and output in real time, and feedback the relevant information to the virtual power plant. The virtual power plant will make necessary adjustments and optimizations to the output plan based on the feedback information to ensure the stable operation of the entire power system and the balance of supply and demand.

[0196] Based on the above, the advantages of the present invention are that the method for constructing the optimal day-ahead scheduling of a virtual power plant containing multiple types of loads based on time-of-use electricity prices is based on the actual power market environment, which can effectively and reasonably improve the income and market competitiveness of virtual power plants and reduce the risks of virtual power plants participating in the market; to a certain extent, the user-side time-of-use electricity prices are linked to different types of users, which helps to improve the income of virtual power plants while reducing the user's electricity costs, and the user's active acceptance is strong; while establishing a day-ahead optimization scheduling model for virtual power plants containing multiple types of loads, pricing strategies for time-of-use electricity prices for three different types of users are also designed, and the two are mutually embedded to obtain the optimal scheduling strategy suitable for virtual power plants and internal loads based on time-of-use electricity prices; after obtaining the day-ahead optimal scheduling strategy for virtual power plants containing multiple types of loads, the peak, valley and flat electricity price data are transmitted to different types of users for load response, and the next day's power generation plan is allocated to each power generation resource for actual power generation. While considering carbon emission rights trading, the internal power generation side and user side of the virtual power plant are mobilized to achieve optimal allocation of internal resources of the virtual power plant and promote the low-carbon economic operation of the virtual power plant.

[0197] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A method for optimizing the day-ahead scheduling of a virtual power plant with multiple types of loads, comprising the following steps: Step 1: power plant investigation and analysis; Step 2: load data collection; Step 3: model building; Step 4: optimization of scheduling strategy; Step 5: plan delivery response; It is characterized by: In the above step 1, based on the centralized spot market, the types of virtual power plants are investigated, which can be divided into demand-side, supply-side and hybrid types according to internal resources; In the above step 2, the peak-valley power demand price elasticity coefficients of different types of users of the virtual power plant operator are calculated, and the users within the virtual power plant are classified into Class A industrial users, Class B industrial companies, commercial users and residential users; Class A industrial users refer to industrial users with high sensitivity to electricity prices and large adjustable load potential, and Class B users have lower responsiveness to electricity prices than Class A industrial users. The elasticity coefficient of the users is calculated through the load data before and after the implementation of the peak-valley time-of-use electricity price; In the above step 3, peak and valley time-of-use electricity prices are implemented for the internal load of the virtual power plant, the peak and valley electricity price response behavior of users is modeled, and the carbon trading mechanism is modeled; In the above step 4, three pricing strategies for time-of-use electricity prices for multiple types of users are proposed. At the same time, the virtual power plant day-ahead optimization scheduling model containing multiple types of loads belongs to a mixed integer nonlinear programming problem. It is solved by optimization software to obtain the optimal scheduling strategy for the virtual power plant day-ahead. In the above step five, the model obtains the optimal time-of-use electricity price data for various types of users within the virtual power plant on the next day, transmits the electricity price information to different users through the communication network, and also obtains the internal power generation resource output plan, which is then distributed to each internal member.

2. The method for optimizing the day-ahead scheduling of a virtual power plant containing multiple types of loads according to claim 1, characterized in that: In the step one, the power plant types are specifically as follows: the demand-side resource type is mainly dominated by the user's own decision-making through transferable and reducible loads based on electricity prices and interruptible loads based on incentives; the supply-side resource type focuses on the peak or valley period of the power grid load, and realizes dynamic regulation of the power supply status by flexibly adjusting the output power of the generator set or optimizing the charging and discharging process of the energy storage equipment to ensure the stable operation of the power grid and the balance of supply and demand; the core of the hybrid resource type is to use advanced energy management systems for intelligent optimization and control, aiming to maximize energy utilization and system economic benefits; the importance of virtual power plants in formulating user peak and valley time-of-use electricity prices and dividing time periods and electricity prices is studied; the demand-side response is divided into price type and incentive type, among which the price type aims to directly transmit market price change signals to terminal power users through the implementation of differentiated electricity price policies, such as time-of-use electricity prices, real-time electricity prices and peak electricity prices. Users adjust their electricity consumption methods according to changes in electricity prices and adopt peak-shifting electricity consumption, which helps to reduce Reduce or transfer part of the peak load, thereby achieving efficient use of energy; price-based demand-side response can, to a certain extent, accurately transmit the price signal in the power supply and demand relationship to users, promote the operation of virtual power plants in a market-oriented manner, enable virtual power plants to more flexibly dispatch and manage decentralized distributed energy resources, and promote efficient use of resources; time-of-use electricity prices belong to the category of price-based demand response, which can guide users to reduce peak loads and fill valleys through the differences in electricity prices in different periods, use electricity scientifically and reduce costs at the same time; the peak and valley time-of-use electricity prices on the user side of virtual power plants should be divided into time periods and the price mechanism should be designed with key considerations on the potential for regulating electricity load characteristics, the need to connect with the spot market, and the degree of acceptability of market players. Time-of-use electricity prices, with their fixed electricity prices within time periods, provide users with clear and predictable price signals. This feature does not require users' metering equipment and response awareness, and has high acceptability and response convenience at the user level, which helps to tap users' response potential and promote efficient use of resources.

3. The method for optimizing the day-ahead scheduling of a virtual power plant containing multiple types of loads according to claim 1, characterized in that: In the step 2, specifically, the model considers that the user's electricity consumption in a certain period will be affected by the comprehensive influence of the electricity price in the current period and the electricity price in other periods, so the peak-valley electricity demand elasticity coefficient expression is: where ε ii is the elastic coefficient, This section shows the electricity price and electricity quantity. Δδ i Respectively represent the amount of change; The electricity price elasticity coefficient ε ij There are two types: self-elasticity and mutual elasticity: In the formula, when j = i, ε ij is the elastic coefficient, j≠i, ε ij is the mutual elastic coefficient; Δδ j They represent the electricity quantity and electricity price in the corresponding time period respectively. Δδ j They represent the changes in the corresponding time periods respectively; The response behavior of users to electricity prices in period n is modeled as follows: Among them, n is the number of time periods, the diagonal elements are the self-elasticity coefficients, and the non-diagonal elements are the cross-elasticity coefficients; various types of load data are collected in real time through intelligent monitoring equipment. Based on this, according to industry characteristics and electricity consumption patterns, users are accurately divided into four categories: Class A industrial users, Class B industrial users, commercial users and residential users; virtual power plant operators use advanced prediction algorithms and big data analysis capabilities to fully analyze the load characteristics of various types of users, such as peak values, valley values, peak-to-valley differences, and load rates, study the changing laws of load curves, and analyze various factors that affect load changes, such as meteorological conditions (temperature, humidity, wind speed, etc.), holidays, special events (power outages, maintenance, etc.), socio-economic factors, etc., and obtain typical daily load curves of various types of loads through the K-means clustering algorithm, and merge them into the total load curve of the virtual power plant operator.

4. The method for optimizing the day-ahead scheduling of a virtual power plant containing multiple types of loads according to claim 3 is characterized by: In the step 2, the peak, valley and average sections of the load curve are divided. Specifically, the peak, valley and average sections of the load are divided by deeply analyzing the peak and valley characteristics of various load curves, and defining the extreme value intervals of each load curve, that is, the highest load point and the lowest load point, and then using the fuzzy boundary gradient membership function mechanism to scientifically divide the peak, valley and average sections of the load; assuming that the set of time points on the typical daily load curve of the user is T = {i1 i2 … i 24 }, the value set corresponding to each period on the curve is L = l1l2 ... l 24 , then the valley time point set corresponding to the valley period is T v , the set of ordinary points corresponding to the ordinary period is T f , the peak time point set corresponding to the peak period is T p , then the valley membership at each time point can be calculated by the fuzzy semi-gradient membership function Sum peak membership Right now; i∈T v i∈T p In order to more accurately calculate the degree of membership at each time point, the present invention replaces l in the formula i It is extended to mean the mean of the load corresponding to time i and time i+1.

5. The method for optimizing the day-ahead scheduling of a virtual power plant containing multiple types of loads according to claim 1, characterized in that: In step 3, the behavior modeling is specifically that when the electricity price is high during a certain period, the user actively responds by reducing the current unnecessary power load or transferring part of the power demand to a period with a more favorable electricity price. Therefore, under the load response mode of 24 periods a day, the load transfer and reduction variables are: Where T p , T f , T v are the peak, flat and valley periods of load, respectively, pf , pv , fv are the interaction transfer coefficients of each period, λ pp , ff , vv are the reduction coefficients for each period, δ p ,δ f ,δ v are the peak, flat and valley prices respectively, k p , k f , k v are the electricity price fluctuation ratios in the corresponding time periods, and k p >0, k f =0, k v <0, so λ ff =0,λ fv =0,λ pf >0,λ pv >0, in addition, due to ε ii < 0, λ pp >0,λ vv <0, we get: Where L' t is the load in period t after execution, L t is the load of the original period t, L p , L f , L v are the average load values ​​for each period.

6. The method for optimizing the day-ahead scheduling of a virtual power plant with multiple types of loads according to claim 5 is characterized by: In the step 3, the VPP components proposed in the present invention include distributed wind power, distributed photovoltaic, micro gas turbine, distributed energy storage and adjustable load; among them, distributed photovoltaic and wind power belong to clean renewable energy, so the carbon emission sources mentioned in the present invention include micro gas turbine, distribution network purchase of electricity (assuming that all are thermal power generation units), and at the same time, according to the national certification voluntary emission reduction mechanism, in addition, according to the national certification voluntary emission reduction mechanism, the environmental protection contribution of wind power and photovoltaic is included in the calculation scope of the free carbon emission quota; the free carbon emission amount of VPP is: Where α is the regional carbon emission quota per unit of electricity, is the power generation of the gas turbine at time t, For wind power, Produce electricity for photovoltaics; The actual carbon emissions of VPP are: Among them, β1 is the carbon emission coefficient of the micro gas turbine, β2 is the carbon emission coefficient of the power purchased by the distribution network, is the power purchased by VPP in the distribution network at time t; The present invention sets up VPP to allow participation in carbon emission rights trading. Under this mode, VPP can flexibly participate in the purchase and sale of carbon emission rights in the carbon trading market, thereby optimizing the configuration and management of carbon assets according to its own operating strategy and market conditions. The carbon trading cost is: where δ c It is the carbon trading price of the day.

7. The method for optimizing the day-ahead scheduling of a virtual power plant with multiple types of loads according to claim 6 is characterized by: In step 3, the specific process of optimizing the scheduling model is as follows: the characteristics of the internal power generation resources of the virtual power plant, the user's electricity cost for: in is the electricity price sold by VPP to user n during period t, is the initial power consumption of user n in period t, is the change in electricity consumption of user n in time period t after responding to the electricity price, and N is the number of user types; Grid transaction costs for: in is the amount of electricity purchased from the distribution network at time t, is the electricity price corresponding to the electricity purchased from the distribution network at time t, is the amount of electricity sold to the distribution network at time t, is the electricity price corresponding to the electricity sold to the distribution network at time t; Gas Turbine Cost for: in is the power generation of the gas turbine at time t, and a, b, and c are their cost coefficients respectively; Energy storage compensation costs for: where δ ev is the storage unit compensation price, is the charging and discharging power of energy storage participating in the electricity market during period t; Wind power penalty costs Photovoltaic penalty costs for: in is the deviation between the actual wind power output and the predicted output during period t, is the penalty price of wind power in period t, is the deviation between the actual photovoltaic output and the predicted output during period t, is the penalty price of photovoltaic power generation during period t; Time-of-use electricity price compensation for users for: in is the load reduction transfer amount of the user in the period n to t, is the compensation price corresponding to user n in period t.

8. The method for optimizing the day-ahead scheduling of a virtual power plant with multiple types of loads according to claim 7 is characterized by: In step 3, in the day-ahead market transaction, the VPP revenue includes the revenue from selling electricity in the day-ahead market and selling electricity to the VPP load, and the VPP cost includes the day-ahead electricity market purchase cost, gas turbine power generation cost, energy storage compensation cost, wind power and photovoltaic penalty cost, carbon trading cost, and user transfer and load reduction compensation cost; the VPP revenue R is: The maximum value of VPP revenue maxR of the objective function of the day-ahead optimization dispatch model of a virtual power plant with multiple types of loads is: in is the revenue of VPP selling electric energy to various types of loads at time t, i.e., the load electricity cost, is the transaction cost of VPP in the day-ahead power market and distribution network at time t, is the gas turbine power generation cost at time t, is the energy storage compensation cost at time t, is the penalty cost of wind power at time t, is the photovoltaic penalty cost at time t, is the carbon trading cost, The time-of-use electricity price compensation cost for users, T is the 24-hour period of the day-ahead dispatch cycle; The constraints of the virtual power plant day-ahead optimal dispatch model with multiple types of loads are constructed. The constraints of the model are power balance constraints, gas turbine constraints, wind and solar output constraints, energy storage constraints, electricity market constraints and user satisfaction constraints. The power balance constraint of the virtual power plant at time t is: in is the amount of electricity purchased from the distribution network at time t, is the amount of electricity sold to the distribution network at time t, is the power generation of the gas turbine at time t, is the charging and discharging power of energy storage participating in the electric energy market during period t, For photovoltaic power output, To contribute to wind power, is the initial power consumption of user n in period t, is the load reduction transfer amount of user n in period t; The gas turbine constraints are: in is the minimum output of the gas turbine, is the maximum output of the gas turbine, is the maximum downward ramp rate of the gas turbine, is the maximum upward climbing rate of the gas turbine; Wind power and photovoltaic output constraints are: in is the predicted wind power output in period t, η1 is the wind power prediction error, is the photovoltaic predicted output during period t, η2 is the photovoltaic prediction error; The energy storage constraint is: Where P es,max is the maximum charging and discharging power of energy storage, is the charging and discharging power of energy storage participating in the electric energy market during period t; when the energy storage device participates in the VPP operation optimization, it must meet the requirement that only charging or discharging is allowed during a certain period of time. The specific constraints are: in is the energy storage charging power during period t, is the energy storage discharge power during period t, U t It is a 0-1 variable, which means that it is 1 if the current period is charged and 0 if it is not charged; in is the initial charge of the energy storage device, is the final state charge of the energy storage device, is the amount of electricity in the energy storage device during period t, E es,min is the minimum capacity of the energy storage device, E es,max is the maximum capacity of the energy storage device, η c is the charging efficiency of the energy storage device, η d is the discharge efficiency of the energy storage device; The electricity market constraints are: in is the maximum power purchase, is the maximum power sold, is the amount of electricity purchased from the distribution network at time t, is the corresponding amount of electricity sold to the distribution network at time t, B t It is a 0-1 variable, indicating the VPP's participation in the electricity market, 1 represents electricity purchase, and 0 represents electricity sales; The user satisfaction constraint is: refers to the electricity satisfaction of user n, It is the lower limit of electricity satisfaction, and its value is 0.

8.

9. The method for optimizing the day-ahead scheduling of a virtual power plant with multiple types of loads according to claim 1, characterized in that: In the step 4, the specific process is to propose three pricing strategies for different types of users to verify the effectiveness of the virtual power plant day-ahead optimization scheduling model; Strategy 1: All users do not use peak and valley time-of-use electricity prices, that is, use flat electricity prices; Strategy 2: Only industrial users adopt peak and valley time-of-use electricity prices; Strategy 3: Industrial, commercial and residential users all adopt peak and valley time-of-use electricity prices according to their own response capabilities; First, Strategy 1 is the strategy currently adopted next to the virtual power plant, which is used as a reference; Secondly, Strategy 2 is a transition strategy between Strategy 1 and Strategy 3. It can be proved that the profit of virtual power plant under Strategy 2 is greater than that under Strategy 1. Finally, Strategy 3 is designed based on the differences in the response capabilities of different types of users to electricity prices. It can fully tap the flexible response potential of different types of users. At the same time, the benefits of the virtual power plant are greater than those of Strategy 2. According to the user load demand, the internal power generation resource parameters (including wind, solar, micro gas turbines, energy storage), time-of-use electricity prices and peak and valley periods in the electricity energy market, and the elasticity coefficients of various types of users are input into the model as input data for the day-ahead optimization scheduling model of the virtual power plant containing multiple types of loads. The model is solved using optimization software to perform optimization and solution. The virtual power plant operator's revenue, user revenue, real-time market revenue of the power sales company, peak electricity prices when different types of users implement time-of-use electricity prices, and the degree of user's active acceptance are obtained under different strategies, and finally the optimal strategy for the day-ahead scheduling of the virtual power plant is determined.

10. The method for optimizing the day-ahead scheduling of a virtual power plant containing multiple types of loads according to claim 1, characterized in that: In the step five, the specific process is to transmit the peak, valley and flat electricity price data to different types of users for load response. First, the prerequisite is prepared; the user needs to have smart devices such as smart meters, load control terminals, data interaction and communication equipment; second, the virtual power plant allocates output plans to power generation resources; the virtual power plant uses advanced information and communication technologies, intelligent control systems, and remote monitoring and dispatching equipment to achieve data exchange and command transmission with power generation resources. Through these technologies, the virtual power plant can send the output plan to the power generation resources in the form of instructions, and collect the operating status, output capacity and other information of the power generation resources in real time, and make timely corrections and adjustments to the behavior that deviates from the plan; third, the power generation resources plan output; after receiving the output plan instruction sent by the virtual power plant, the power generation resources will adjust their own operating status and output level according to the requirements in the instruction. In the process of executing the planned output, the power generation resources will monitor their own operating status and output in real time, and feedback the relevant information to the virtual power plant. The virtual power plant will make necessary adjustments and optimizations to the output plan based on the feedback information to ensure the stable operation and supply and demand balance of the entire power system.

Images