Micro-grid two-stage optimization scheduling method based on fuzzy opportunity constraint and master-slave game

Abstract

The invention discloses a microgrid two-stage optimization scheduling method based on fuzzy opportunity constraint and a master-slave game. The microgrid two-stage optimization scheduling method comprises the following steps: step 1, establishing a microgrid mathematical model; 2, in a day-ahead scheduling stage, a micro-grid master-slave game double-layer optimization model based on an excitation type demand response mechanism is established, and game subjects in the model are a power distribution network operator and a micro-grid operator respectively; 3, in an intra-day scheduling stage, proposing an intra-day rolling optimization scheduling strategy based on model prediction control, and realizing optimization adjustment of a day-ahead scheduling plan; 4, depicting the uncertainty of wind and light output and load under different time scales by using fuzzy opportunity constraints, converting the determinacy constraints in the system into fuzzy condition constraints, and meanwhile, performing clearness processing on fuzzy parameters by using a trapezoidal function; 5, in a day-ahead scheduling stage, iteratively solving the upper layer optimization model and the lower layer optimization model by using a particle swarm optimization algorithm and a CPLEX solver; in the intra-day scheduling stage, a CPLEX solver is adopted for iterative solution.

Description

Technical Field

[0001] The invention relates to the technical field of power system optimization dispatching, and in particular to a two-stage optimization dispatching method for a microgrid based on fuzzy chance constraints and master-slave game. Background Art

[0002] In response to the deteriorating climate, the world is vigorously developing new power systems based on renewable energy such as wind power and photovoltaics. Microgrids can flexibly integrate various types of renewable energy, and can achieve efficient use and optimized management of energy. At the same time, the intermittent and random nature of wind power and photovoltaic power generation always threatens the stable operation of the power grid, bringing difficulties and challenges to the optimized operation of microgrids.

[0003] At present, there have been some studies on the optimization and dispatching of microgrids. "Research on Typical Daily Dispatching of Multi-energy Parks Based on Improved DDPG" in Volume 46, Issue 5 of "Power System Technology" solves the model with the goal of optimizing the economic efficiency of microgrid operation, which effectively improves the operation efficiency of the system. "Optimal Operation Strategy of Micro-integrated Energy System Considering the Heat and Power Cogeneration / Supply Characteristics of AA-CAES Device" in Volume 46, Issue 2 of "High Voltage Technology" proposes a combined heat and power system based on compressed air energy storage to improve the system economy and renewable energy absorption capacity. However, considering the operation economy of the system alone can no longer meet the requirements of the new power system at this stage. In view of the diversification of load demand, the introduction of demand response can not only achieve energy complementarity but also reduce the energy supply pressure of the power system. For this reason, "Optimal Dispatching of Integrated Energy System Considering the Complementarity of Carbon Capture and Comprehensive Demand Response" in Volume 42, Issue 7 of "Electric Power Automation Equipment" combines the step-by-step carbon trading mechanism and comprehensive demand response to effectively improve the system operation economy and environmental protection. The above research provides an important theoretical basis for the optimal dispatching of power systems, but does not consider the flexible interaction of power systems in participating in the market. The 48th issue of Automation of Electric Power Systems, “Low-carbon Optimal Dispatch of Integrated Energy Systems Considering Low-carbon Demand Response and Master-slave Game” is based on incentive-based low-carbon demand response and carbon trading, taking into account the game process of internal electricity prices in the system, effectively improving the economic efficiency of system operation. The 43rd issue of Automation of Electric Power Systems, “Interactive Optimal Operation Method of Integrated Energy Systems in Industrial Parks Based on Master-slave Game” takes the IES system in industrial parks as the research object, proposes a Stackelberg game model and combines the characteristics of thermal-electric coupling loads, effectively improving the economic stability of system operation.

[0004] The above studies are all based on accurate prediction data of sources and loads, while ignoring their uncertainty factors. In view of the uncertainty of renewable energy output on the source side and multi-energy load on the demand side, most of the current literature uses analysis methods such as probabilistic scenario method, robust optimization method and chance constraint programming method. "Automation of Electric Power Systems" Vol. 47 No. 20 "Microgrid Typical Scenario Generation Method Based on Deep Time Clustering" proposes a typical scenario generation method of combined neural network for the uncertainty of renewable energy and load, which can fully explore the complex characteristics of the data. "Power System Protection and Control" Vol. 51 No. 6 "Robust Optimal Dispatch of Integrated Energy Systems Containing Electricity-to-Hydrogen Considering Flexible Supply and Demand Balance" proposes a robust optimization model for the electric-to-hydrogen IES system, while taking into account the supply and demand balance of the uncertain load side of renewable energy.

[0005] The above methods mostly rely on historical data or probability distribution. If the historical data is insufficient or of low quality, the scenarios generated may not be accurate enough, affecting the reliability of the results. The fuzzy chance constraint method can analyze uncertainty by fuzzifying the predicted value and establishing a membership function. At the same time, a clear class equivalence method is used to transform the fuzzy chance constraint into a deterministic constraint to improve the efficiency of solving the power system.

[0006] In addition, the uncertainty of wind and solar power output and load demand will also cause their prediction errors to increase with the increase of time scale, and the day-ahead dispatch plan under a longer time scale is difficult to meet the multi-energy coordinated dispatch of the power system. Therefore, based on the analysis of the uncertainty at both the source and the load, it is becoming increasingly important to study the multi-time scale dispatch problem of the power system. Summary of the invention

[0007] The technical problem to be solved by the present invention is to provide an overhead LNG pipeline support structure with settlement compensation adjustment, which can increase the height of the pipeline fixing device to compensate for the settlement height when settlement occurs at the support.

[0008] In order to solve the above technical problems, the technical solution adopted by the present invention is: a two-stage optimization scheduling method of a microgrid based on fuzzy chance constraints and master-slave game, comprising the following steps: Step 1: Establish a mathematical model of the microgrid; Step 2: In the day-ahead dispatch stage, a two-layer optimization model of microgrid master-slave game based on incentive-based demand response mechanism is established. The game subjects in the model are distribution network operators (DNO) and microgrid operators (MO). Step 3: In the intraday scheduling stage, an intraday rolling optimization scheduling strategy based on model predictive control is proposed to achieve the optimization adjustment of the day-ahead scheduling plan; Step 4: Use fuzzy opportunity constraints to characterize the uncertainty of wind and solar power output and load at different time scales, transform the deterministic constraints in the system into fuzzy conditional constraints, and use trapezoidal functions to clarify the fuzzy parameters; Step 5: In the day-ahead scheduling stage, the particle swarm algorithm and CPLEX solver are used to iteratively solve the upper optimization model and the lower optimization model respectively; the CPLEX solver is used to iteratively solve the intraday scheduling stage.

[0009] Preferably, the mathematical model of the microgrid in step one includes a fan output model, a photovoltaic output model, a combined heat and power unit, an electric boiler, a gas boiler, a battery, and a heat storage tank mathematical model.

[0010] Preferably, each model is as follows: (1) Fan output model ; In the formula, is the output power of the fan at time t; is the wind speed at time t; , , They are the fan’s cut-in speed, rated speed and cut-out speed respectively; is the rated output power of the fan; (2) Photovoltaic output model ; In the formula, is the photovoltaic output power at time t; is the rated power of the PV; and are the reference irradiance and reference temperature respectively; and are the solar irradiance and ambient temperature at time t, respectively; is the power temperature coefficient, usually taken as 0.0047°C -1 ; (3) Combined heat and power unit model ; Where: are the electric and thermal output power of the cogeneration unit respectively; is the gas turbine electrical conversion efficiency; is the electric-to-heat ratio of the CHP unit; is the heat conversion efficiency of the CHP unit; The total power output of the micro gas turbine; (4) Electric boiler model ; Where: are the heat generation and power consumption of the electric boiler at time t respectively; is the thermoelectric conversion efficiency; (5) Gas boiler model ; Where: is the heating power of the gas boiler at time t; is the heat generation efficiency of the gas boiler; Power consumption for natural gas; (6) Battery model ; Where: The battery storage capacity at time t; is the battery charge state at time t; is the rated capacity of the battery; are the charging and discharging power of the battery at time t respectively; are the charge and discharge efficiency of the battery respectively; (7) Heat storage tank model ; Where: The heat storage tank has a heat storage capacity at time t; is the heat storage state of the heat storage tank at time t; is the rated capacity of the heat storage tank; are the heat storage and release power of the heat storage tank at time t respectively; are the heat storage and release efficiencies of the heat storage tank respectively.

[0011] Preferably, in step 2, the day-ahead scheduling plan cycle is set to 24 hours, and the time scale is set to 1 hour. Based on the power forecast results of wind power, photovoltaic power and load, a day-ahead master-slave game two-layer optimization scheduling model based on incentive-based demand response is established. The optimization target of the upper model is to maximize the daily operating revenue of the distribution network operator, and the optimization target of the lower model is to minimize the operating cost of the microgrid operator.

[0012] Preferably, the upper layer objective function is as follows: The goal of the distribution network operator is to maximize daily revenue. The optimization problem can be expressed as: ; In the formula, is the power purchase and sale cost of the distribution network operator under different load types; is the demand response subsidy cost; k represents the load type; Equipment operation and maintenance costs; The cost of pollutant emissions; For the tiered carbon trading costs; The cost of purchasing gas; Tiered carbon trading costs The daily carbon emissions trading volume in the carbon market is: ; Where: It is the daily carbon emission quota; is the actual daily carbon emissions; The formula for calculating the tiered carbon trading cost is: ; Where: It is the base price for carbon trading; For carbon trading quotas; is the carbon price growth rate; (2) Demand response subsidies ; Where: Subsidize prices for users; It represents the total power of the kth type of load participating in demand response at time t.

[0013] Preferably, the lower layer objective function is as follows: The optimization goal of the follower is to minimize the daily energy cost, which can be expressed as: The day-ahead objective function is: ; In the formula, Regulation costs of changing electricity consumption schedules for microgrids to participate in incentivized demand response.

[0014] Preferably, in step 3, the intraday rolling scheduling plan cycle is set to 4h, the time scale is set to 15min, and 96 rolling optimizations are performed in one day. The intraday optimization scheduling objective function is established with the goal of minimizing the deviation between the output vector of the intraday plan and the day-ahead optimization scheduling plan; specifically as follows: (1) Intraday objective function ; Where: and are the output vector and the reference vector respectively, is the output value of each device in the day-ahead scheduling plan; Q is the error coefficient matrix; R is the control coefficient matrix; m is the number of control time domain periods; U is the control vector; (2) Feedback correction link ; Where: To optimize the initial power of the unit; is the actual power of the unit.

[0015] Preferably, in step 4, fuzzy opportunity constraints are used to characterize the uncertainty of wind and solar power output and load at different time scales, and the uncertainty constraints in the system are converted into fuzzy conditional constraints with confidence, and the fuzzy parameters are clarified by using trapezoidal functions; specifically as follows: ; In the formula, is the decision variable, is the fuzzy vector, is the objective function, is the constraint function.

[0016] Considering that the fuzzy variables in the constraints will cause the deterministic constraints to become uncertain, the confidence factor is introduced. , transforming uncertainty constraints into chance constraints.

[0017] ; In the formula, for credibility of occurrence; The fuzzy variables of wind power generation, photovoltaic power generation and various load forecasts can be expressed by trapezoidal functions. The details are as follows: ; In the formula, is the fuzzy parameter membership function, (i=1, 2, 3, 4) is the membership function of the trapezoidal function. When the trapezoidal function is converted into trigonometric function. The relationship with wind power generation, photovoltaic power generation and various load forecasts is: ; In the formula, The proportional coefficient range is ; Predicted values ​​for wind power, photovoltaic power and various loads.

[0018] The present invention provides a two-stage optimization scheduling method for a microgrid based on fuzzy chance constraints and master-slave game, and the specific beneficial effects are as follows: 1. Taking microgrid as the research object, it can flexibly integrate various distributed energy sources and reduce the operating cost of the system.

[0019] 2. Establish distribution network operators (DNO) and microgrid operators (MO) as entities with different interests, and their decisions will affect each other. In the day-ahead scheduling stage, the present invention uses a master-slave game process to allow both parties to fully consider the impact of each other's decisions on their own interests, thereby formulating a better electric heat price and maximizing the overall benefits of the system. At the same time, the load-side incentive demand response is introduced to guide users to adjust their electricity demand, so that the distribution network operator (DNO) can better achieve effective regulation of the grid load and improve the flexibility and stability of the system power supply. 3. In order to solve the problem of decreased scheduling plan accuracy caused by a single optimization time scale, the present invention combines the model predictive control rolling optimization strategy in the intraday scheduling stage to achieve optimized adjustment of the day-ahead scheduling plan to improve the stability of system operation.

[0020] 4. Taking into account the impact of source-load uncertainty on system scheduling, the present invention introduces fuzzy chance constraints to transform deterministic constraints into uncertainty constraints containing fuzzy variables, so that the economic efficiency of system operation can be analyzed at different confidence levels during the decision-making process, which is conducive to formulating more complete response strategies and further improving the flexibility and stability of system operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The present invention will be further described below in conjunction with the accompanying drawings and embodiments: Figure 1 is a flow chart of the steps of the present invention; Figure 2 It is a diagram of multi-time scale scheduling steps; Figure 3a This is the result diagram of the demand response quantity of the electric load; Figure 3b The result diagram of demand response quantity for heat load participation; Figure 4 Optimal pricing map for distribution network operators; Figure 5a This is the result diagram of the electricity balance of the day-ahead optimized dispatch; Figure 5b This is the heat balance result diagram of the day-ahead optimization scheduling; Figure 6a This is the result diagram of the intraday optimized dispatching electricity balance; Figure 6b This is the thermal balance result diagram for intraday optimization scheduling; Figure 7 The graph is a relationship between different confidence levels and the revenue of distribution network operators; Figure 8 It is a curve diagram showing the relationship between fuzzy membership and the profit of distribution network operators. DETAILED DESCRIPTION

[0022] The present invention comprises the following steps: Step 1: Establish mathematical models of combined heat and power units, electric boilers, gas boilers, batteries and heat storage tanks. The details are as follows: System flow chart see Figure 1 .

[0023] (1) Fan output model ; In the formula, is the output power of the fan at time t; is the wind speed at time t; , , They are the fan’s cut-in speed, rated speed and cut-out speed respectively; is the rated output power of the fan.

[0024] (2) Photovoltaic output model ; In the formula, is the photovoltaic output power at time t; is the rated power of the PV; and are the reference irradiance and reference temperature respectively; and are the solar irradiance and ambient temperature at time t, respectively; is the power temperature coefficient, usually taken as 0.0047°C-1.

[0025] (3) Combined heat and power unit model ; Where: are the electric and thermal output power of the cogeneration unit respectively; is the gas turbine electrical conversion efficiency; is the electric-to-heat ratio of the CHP unit; is the heat conversion efficiency of the CHP unit; The total power output of the micro gas turbine.

[0026] (4) Electric boiler model ; Where: are the heat generation and power consumption of the electric boiler at time t respectively; is the thermoelectric conversion efficiency.

[0027] (5) Gas boiler model ; Where: is the heating power of the gas boiler at time t; is the heat generation efficiency of the gas boiler; Power consumption for natural gas.

[0028] (6) Battery model ; Where: The battery storage capacity at time t; is the battery charge state at time t; is the rated capacity of the battery; are the charging and discharging power of the battery at time t respectively; are the charge and discharge efficiencies of the battery respectively.

[0029] (7) Heat storage tank model ; Where: The heat storage tank has a heat storage capacity at time t; is the heat storage state of the heat storage tank at time t; is the rated capacity of the heat storage tank; are the heat storage and release power of the heat storage tank at time t respectively; are the heat storage and release efficiencies of the heat storage tank respectively.

[0030] The multi-time scale scheduling steps are shown in Figure 2 .

[0031] Step 2: Set the day-ahead dispatch plan cycle to 24 hours and the time scale to 1 hour. Based on the power forecast results of wind power, photovoltaic power and load, a day-ahead master-slave game two-layer optimization dispatch model based on incentive-based demand response is established. The optimization goal of the upper model is to maximize the daily operating revenue of the distribution network operator, and the optimization goal of the lower model is to minimize the operating cost of the microgrid operator. The details are as follows: (a) Upper objective function The goal of the distribution network operator is to maximize daily revenue. The optimization problem can be expressed as: ; In the formula, is the power purchase and sale cost of the distribution network operator under different load types; is the demand response subsidy cost; k represents the load type; Equipment operation and maintenance costs; The cost of pollutant emissions; For the tiered carbon trading costs; The cost of purchasing gas.

[0032] (1) Tiered carbon trading mechanism The daily carbon emissions trading volume in the carbon market is:

[0033] Where: It is the daily carbon emission quota; is the actual daily carbon emissions; The formula for calculating the tiered carbon trading cost is: ; Where: It is the base price for carbon trading; For carbon trading quotas; is the carbon price growth rate; (2) Incentive Demand Response Subsidy ; Where: Subsidize prices for users; It represents the total power of the kth type of load participating in demand response at time t.

[0034] (3) Operation and maintenance costs of each device ; Where: They are the operation and maintenance costs of batteries, micro gas turbines, waste heat boilers, electric boilers, gas boilers and heat storage tank equipment.

[0035] (4) Microgrid electric heat purchase and sale costs ; Where: and are the purchase and sale prices of the kth type of load respectively; and They are the purchased and sold power of the kth type of load at time t respectively.

[0036] (5) Pollutant emission costs ; Where: are the pollutant emission cost coefficients of CHP units, gas boilers and microgrid interactions; They are the pollutant emission coefficients of CHP units, gas boilers and microgrid interactions.

[0037] (6) Gas purchase cost ; Where: is the natural gas price; a, b, c are the combustion coefficients respectively.

[0038] (b) Lower layer objective function The optimization goal of the follower is to minimize the daily energy cost, which can be expressed as: The day-ahead objective function is: ; In the formula, Regulation costs of changing electricity consumption schedules for microgrids to participate in incentivized demand response.

[0039] The demand response regulation cost is: ; Where m and n are the regulation cost coefficients when the microgrid participates in demand response.

[0040] Specifically, in the day-ahead optimization scheduling model, the constraints mainly include the following aspects: a) Electric heating power balance power ; ; Where: The photovoltaic power prediction at time t; Predict the power of wind turbines at time t; The electric power before participating in demand response; is the thermal power before participating in demand response.

[0041] b) Battery output constraints ; Where: are the maximum charging and discharging power of the battery at time t; , They are 0-1 state variables, 1 indicates the battery is in a charging state, otherwise it is in a discharging state; are the maximum and minimum state of charge of the battery at time t respectively.

[0042] c) Heat storage tank output constraints ; Where: , They are the maximum values ​​of heat storage and release power of the heat storage tank at time t respectively; They are 0-1 state variables, 1 means the heat storage tank is in the heat storage state, otherwise it is in the heat release state; are the maximum and minimum values ​​of the heat storage state of the heat storage tank at time t respectively.

[0043] d) Tie line power constraints ; Where: is the maximum interactive power of the tie line; They are 0-1 state variables, 1 means the electric heat quantity is purchased, otherwise it is sold; Represents the interaction power of the kth type load tie line.

[0044] e) Demand-side response constraints ; Where: is the power after participating in demand response.

[0045] f) User satisfaction constraints In order to balance the relationship between the economic benefits of the power system and user experience, the Integrated Consumer Satisfaction Index (ICSI) is used to characterize user satisfaction with electricity consumption.

[0046] ; Where: It represents the minimum comprehensive consumer satisfaction index; Represents the total power of the electrical load at time t.

[0047] Step 3: The intraday rolling scheduling plan cycle is set to 4 hours, the time scale is set to 15 minutes, and 96 rolling optimizations are performed in one day. Based on the ultra-short-term wind and solar power forecast results, the intraday optimization scheduling objective function is established with the goal of minimizing the deviation between the output vector of the intraday plan and the day-ahead optimization scheduling plan. The details are as follows: (1) Intraday objective function Based on the rolling optimization scheduling strategy of MPC, the change of electric heat charging and discharging of batteries and heat storage tanks in the microgrid is used as the control vector. By adjusting the size of the control vector, the deviation between the output vector of the intraday plan and the day-ahead optimization scheduling plan is minimized. The established objective function is as follows: ; Specifically, in the above formula: ; Where: and are the output vector and the reference vector respectively, is the output value of each device in the day-ahead scheduling plan; Q is the error coefficient matrix; R is the control coefficient matrix; m is the number of control time domain periods; is the control vector.

[0048] (2) Feedback correction link In order to further reduce the impact of errors caused by two-stage forecasting, a feedback correction mechanism is added to the intraday rolling optimization scheduling to achieve closed-loop control of the system. The feedback correction link is as follows: ; Where: To optimize the initial power of the unit; is the actual power of the unit.

[0049] In the intraday optimization scheduling model, the constraints are the same as those in the day-ahead optimization scheduling model.

[0050] Step 4: Use fuzzy opportunity constraints to characterize the uncertainty of wind and solar power output and load at different time scales, transform the uncertainty constraints in the system into fuzzy conditional constraints with confidence, and use trapezoidal functions to clarify the fuzzy parameters. The details are as follows: ; In the formula, is the decision variable, is the fuzzy vector, is the objective function, is the constraint function.

[0051] Considering that the fuzzy variables in the constraints will cause the deterministic constraints to become uncertain, the confidence factor is introduced. , transforming uncertainty constraints into chance constraints.

[0052] ; In the formula, for The credibility of what happened.

[0053] Therefore, the electrothermal balance constraint under a certain confidence level is: ; ; The fuzzy variables of wind power generation, photovoltaic power generation and various load forecasts can be expressed by trapezoidal functions. The details are as follows: ; In the formula, is the fuzzy parameter membership function, (i=1, 2, 3, 4) is the membership function of the trapezoidal function. When the trapezoidal function is converted into trigonometric function. The relationship with wind power generation, photovoltaic power generation and various load forecasts is: ; In the formula, The proportional coefficient range is ; Predicted values ​​for wind power, photovoltaic power and various loads.

[0054] In the fuzzy opportunity constraint output result, the clear equivalent of the power balance constraint is: ; ; In the formula, , , as well as are the membership functions of electric and heating loads respectively; , , as well as are the membership parameters of the wind power and photovoltaic power generation output prediction values ​​respectively.

[0055] Step 5: In the day-ahead scheduling stage, the particle swarm algorithm and CPLEX solver are used to iteratively solve the upper optimization model and the lower optimization model respectively; the CPLEX solver is used to iteratively solve the intraday scheduling stage.

[0056] The effectiveness of the present invention is verified by taking a typical microgrid in China as an example: The demand response amount of electric and heating loads is shown in Figure 3a , Figure 3b .

[0057] pass Figure 3a , Figure 3b It can be found that the load side reduces or transfers some of the electric and thermal loads during peak hours to valley and normal hours by participating in demand response, thus achieving the "peak shaving and valley filling" of the load. By encouraging the user side to participate in demand response, it can not only effectively reduce the operating cost of the microgrid, but also better cope with the impact of source and load uncertainty.

[0058] Figure 4 The optimal pricing result of the distribution network operator is obtained by solving the day-ahead master-slave game two-level optimization model.

[0059] The day-ahead scheduling results are as follows: Figure 5a , Figure 5b As shown in the power balance diagram, the amount of electricity purchased fluctuates greatly in different time periods. In the period from 1:00 to 6:00, the photovoltaic output is 0, and the wind turbine provides most of the load demand for users. Figure 4The optimal pricing results of the distribution network operator's electric heat obtained by iteration show that the electricity price is the lowest at this time, and the system fills the remaining electric heat load demand by purchasing a certain amount of electricity from the superior power grid. During the period of 6:00-10:00, the photovoltaic output gradually increases, and the electric load demand gradually increases. The system also needs to discharge the battery to meet the user's electricity demand. During the period of 10:00-15:00, the photovoltaic output gradually reaches the maximum value. At this time, in addition to using the power generation of each unit, the system also purchases electricity from the superior power grid to meet the electricity demand and stores the excess electricity for nighttime backup. In the period of 15:00-21:00, the photovoltaic output gradually decreases. The system cannot meet the electricity demand only through wind power generation and power generation of each unit. At this time, the electricity price is low, so the system purchases electricity from the superior power grid to fill the user's electricity shortage. In the period of 21:00-1:00, the electricity demand is low and the heat load demand is high. The system fills the remaining electric heat load demand by purchasing electricity and using electricity-to-heat equipment.

[0060] It can be seen from the thermal power balance diagram that in the 00:00-6:00 time period, the heat load is high, and the system mainly uses WHB and electric boilers for heating, and then the heat storage tank releases heat to fill the heat demand. In the 06:00-12:00 time period, the heat load output gradually decreases, and the system mainly uses WHB and electric boilers for heating. In the 12:00-18:00 time period, users' heat demand is not large. The system uses WHB and electric boilers for heating and stores excess heat in the heat storage tank for use at night. In the 18:00-24:00 time period, the system mainly uses WHB and electric boilers for heating, and the heat storage tank releases heat after 19:00 to fill the heat load shortfall. In addition, by Figure 4 From the optimal heat price obtained, it can be found that the heat price does not fluctuate much throughout the day. Taking into account the electric-thermal coupling characteristics within the system, while ensuring optimal economic operation, the system purchases a certain amount of heat from the superior power grid in each time period to meet the heat demand.

[0061] The intraday scheduling results are as follows: Figure 6a , Figure 6b By comparing the day-ahead intraday dispatch plan, it can be seen that in the intraday dispatch stage, the dispatch results are more refined. According to the day-ahead dispatch plan, the system can reduce the impact of wind and solar forecast errors and load fluctuations by adjusting the output of various equipment such as interconnection lines, batteries, and heat storage tanks, thereby improving the stability of system operation.

[0062] Considering the uncertainty of source load, the parameters of the trapezoidal function are set as shown in Table 1. The confidence level interval is [0.8, 1]. The relationship curve between different confidence levels and the revenue of distribution network operators is shown in Figure 7 As shown. Figure 7It can be seen that as the confidence level continues to increase, the income of distribution network operators is gradually decreasing. At this time, the system improves the stability of system operation by giving up certain economic benefits. When the confidence level is in the interval [0.9,1], the system has a certain stability, and the income of distribution network operators has not decreased significantly, so 0.9 is selected as the optimal confidence level.

[0063] Table 1 Trapezoid fuzzy membership parameters

[0064] In view of the uncertainty of source and load, different degrees of fuzzy membership are used for analysis. Table 2 shows the fuzzy parameters of wind and solar power output and load under five fuzzy memberships, W1~W4 and W5~W8 respectively.

[0065] Table 2 Trapezoidal fuzzy membership parameters

[0066] Depend on Figure 8 It can be seen that under the specified confidence level, the revenue of the distribution network operator decreases as the source-load ambiguity increases, which shows that the system can ensure the stable operation of the system in a complex environment by sacrificing part of the economy.

[0067] The present invention proposes a two-stage optimization scheduling method for a microgrid based on fuzzy chance constraints and master-slave game. First, in the day-ahead scheduling stage, a two-layer optimization model of microgrid master-slave game based on incentive-based demand response mechanism is established, and the uncertainty of renewable energy and multi-energy load forecast values ​​is analyzed based on fuzzy chance constraints. Secondly, in the intraday stage, an intraday rolling optimization scheduling strategy based on model predictive control is proposed to optimize and adjust the day-ahead scheduling plan to improve the stability of system operation. Finally, the two-layer optimization model of the day-ahead scheduling stage is combined with the particle swarm optimization algorithm and the CPLEX solution period for iterative solution. The method proposed in the present invention can effectively improve the economy, environmental protection and robustness of system operation.

[0068] The above embodiments are only preferred technical solutions of the present invention and should not be regarded as limiting the present invention. The protection scope of the present invention shall be the technical solutions recorded in the claims, including equivalent replacement solutions of the technical features in the technical solutions recorded in the claims. That is, equivalent replacement improvements within this scope are also within the protection scope of the present invention.

Claims

1. A two-stage optimization scheduling method for microgrids based on fuzzy chance constraints and master-slave game, characterized in that: The following steps are involved: Step 1: Establish a mathematical model of the microgrid; Step 2: In the day-ahead dispatch stage, a two-layer optimization model of microgrid master-slave game based on the incentive-based demand response mechanism is established. The game subjects in the model are the distribution network operator and the microgrid operator. Step 3: In the intraday scheduling stage, an intraday rolling optimization scheduling strategy based on model predictive control is proposed to achieve the optimization adjustment of the day-ahead scheduling plan; Step 4: Use fuzzy opportunity constraints to characterize the uncertainty of wind and solar power output and load at different time scales, transform the deterministic constraints in the system into fuzzy conditional constraints, and use trapezoidal functions to clarify the fuzzy parameters; Step 5: In the day-ahead scheduling stage, the particle swarm algorithm and CPLEX solver are used to iteratively solve the upper optimization model and the lower optimization model respectively; the CPLEX solver is used to iteratively solve the intraday scheduling stage.

2. According to claim 1, a two-stage optimization scheduling method for microgrids based on fuzzy chance constraints and master-slave game, characterized in that: The mathematical model of the microgrid in step one includes a fan output model, a photovoltaic output model, a combined heat and power unit, an electric boiler, a gas boiler, a battery, and a heat storage tank mathematical model.

3. According to claim 2, a two-stage optimization scheduling method for microgrids based on fuzzy chance constraints and master-slave game, characterized in that: The models are as follows: (1) Fan output model ； In the formula, is the output power of the fan at time t; is the wind speed at time t; , , They are the fan’s cut-in speed, rated speed and cut-out speed respectively; is the rated output power of the fan; (2) Photovoltaic output model ； In the formula, is the photovoltaic output power at time t; is the rated power of the PV; and are the reference irradiance and reference temperature respectively; and are the solar irradiance and ambient temperature at time t, respectively; is the power temperature coefficient; (3) Combined heat and power unit model ； Where: are the electric and thermal output power of the cogeneration unit respectively; is the gas turbine electrical conversion efficiency; is the electric-to-heat ratio of the CHP unit; is the heat conversion efficiency of the CHP unit; The total power output of the micro gas turbine; (4) Electric boiler model ； Where: are the heat generation and power consumption of the electric boiler at time t respectively; is the thermoelectric conversion efficiency; (5) Gas boiler model ； Where: is the heating power of the gas boiler at time t; is the heat generation efficiency of the gas boiler; Power consumption for natural gas; (6) Battery model ； Where: The battery storage capacity at time t; is the battery charge state at time t; is the rated capacity of the battery; are the charging and discharging power of the battery at time t respectively; are the charge and discharge efficiency of the battery respectively; (7) Heat storage tank model ； Where: The heat storage tank has a heat storage capacity at time t; is the heat storage state of the heat storage tank at time t; is the rated capacity of the heat storage tank; are the heat storage and release power of the heat storage tank at time t respectively; are the heat storage and release efficiencies of the heat storage tank respectively.

4. According to claim 1, a two-stage optimization scheduling method for microgrids based on fuzzy chance constraints and master-slave game, characterized in that: In the step 2, based on the power forecast results of wind power, photovoltaic power and load, a day-ahead master-slave game two-layer optimization scheduling model based on incentive-based demand response is established. The optimization goal of the upper model is to maximize the daily operating profit of the distribution network operator, and the optimization goal of the lower model is to minimize the operating cost of the microgrid operator.

5. According to claim 4, a two-stage optimization scheduling method for microgrids based on fuzzy chance constraints and master-slave game, characterized in that: The upper layer objective function is as follows: The goal of the distribution network operator is to maximize daily revenue. The optimization problem can be expressed as: ； In the formula, is the power purchase and sale cost of the distribution network operator under different load types; is the demand response subsidy cost; k represents the load type; Equipment operation and maintenance costs; The cost of pollutant emissions; For the tiered carbon trading costs; The cost of purchasing gas; Tiered carbon trading costs The daily carbon emissions trading volume in the carbon market is: ； Where: It is the daily carbon emission quota; is the actual daily carbon emissions; The formula for calculating the tiered carbon trading cost is: ； Where: It is the base price for carbon trading; For carbon trading quotas; is the carbon price growth rate; (2) Demand response subsidies ； Where: Subsidize prices for users; It represents the total power of the kth type of load participating in demand response at time t.

6. According to claim 4, a two-stage optimization scheduling method for microgrids based on fuzzy chance constraints and master-slave game, characterized in that: The lower layer objective function is as follows: The optimization goal of the follower is to minimize the daily energy cost, which can be expressed as: The day-ahead objective function is: ； In the formula, Regulation costs of changing electricity consumption schedules for microgrids to participate in incentivized demand response.

7. According to claim 1, a two-stage optimization scheduling method for microgrids based on fuzzy chance constraints and master-slave game, characterized in that: In step 3, the intraday optimization scheduling objective function is established with the goal of minimizing the deviation between the output vector of the intraday plan and the day-ahead optimization scheduling plan; the details are as follows: (1) Intraday objective function ； Where: and are the output vector and the reference vector respectively, is the output value of each device in the day-ahead scheduling plan; Q is the error coefficient matrix; R is the control coefficient matrix; m is the number of control time domain periods; U is the control vector; (2) Feedback correction link ； Where: To optimize the initial power of the unit; is the actual power of the unit.

8. According to claim 1, a two-stage optimization scheduling method for microgrids based on fuzzy chance constraints and master-slave game, characterized in that: In the fourth step, fuzzy opportunity constraints are used to characterize the uncertainty of wind and solar power output and load at different time scales, and the uncertainty constraints in the system are converted into fuzzy conditional constraints with confidence. At the same time, trapezoidal functions are used to clarify the fuzzy parameters; specifically, as follows: ； In the formula, is the decision variable, is the fuzzy vector, is the objective function, is the constraint function. Considering that the fuzzy variables in the constraints will cause the deterministic constraints to become uncertain, the confidence factor is introduced. , transforming uncertainty constraints into chance constraints. ； In the formula, for credibility of occurrence; The fuzzy variables of wind power generation, photovoltaic power generation and various load forecasts can be expressed by trapezoidal functions. The details are as follows: ； In the formula, is the fuzzy parameter membership function, (i=1, 2, 3, 4) is the membership function of the trapezoidal function. When the trapezoidal function is converted into trigonometric function. The relationship with wind power generation, photovoltaic power generation and various load forecasts is: ； In the formula, The proportional coefficient range is ; Predicted values ​​for wind power, photovoltaic power and various loads.

Images