An optimization method for connecting distributed energy to microgrids

Abstract

The present invention discloses an optimization method for connecting distributed energy to a microgrid, including inputting the meteorological forecast data of the day into the working model of each power generation equipment to obtain the output forecast data of each power generation equipment; performing similar collaborative clustering based on the load level matching strategy on the load data of the day to obtain the load level type and the load removal strategy of different levels in different time periods; establishing an operation mode strategy according to the output forecast data of each power generation equipment, the load level type and the load removal strategy of different levels in different time periods, and establishing a microgrid daily scheduling optimization model according to the working model of each distributed unit, the operation mode strategy, the total cost model of microgrid operation and various constraints; solving the microgrid daily scheduling optimization model by improving the particle swarm algorithm to obtain the microgrid daily optimization scheduling plan, and connecting the distributed energy to the microgrid according to the microgrid daily optimization scheduling plan. This embodiment achieves the improvement of the power supply reliability and operation stability of the microgrid.

Description

Technical Field

[0001] The present invention relates to the field of power grid dispatching, and in particular to an optimization method for accessing distributed energy to a microgrid. Background Art

[0002] The traditional centralized energy supply system transmits various energies to many users in a large area through special transmission facilities, while the distributed energy system directly faces users and supplies energy locally according to their needs. Microgrids combine different distributed energy sources to improve energy utilization. With the introduction of various distributed energy sources, the original operation mode of microgrids has been broken.

[0003] When connecting distributed energy to microgrids, the uncertainty and volatility of energy sources such as photovoltaics and wind turbines also interfere with the operation of various devices, making the stability of microgrid operation unguaranteed. In addition, traditional dispatching strategies that perform single dispatching operations on loads or mismatched output of equipment will cause a large amount of power and heat losses in the microgrid, reducing the reliability of the system. Therefore, how to effectively manage the output of each distributed power source within the microgrid and maximize the economic and environmental benefits of the microgrid has been a hot topic and difficult problem in recent years. Summary of the invention

[0004] The present invention provides an operation optimization method for connecting distributed energy to a microgrid, which realizes a scheduling scheme for optimizing the distributed energy connected to the microgrid, ensures the power supply reliability of the microgrid, and improves the stability of the microgrid operation under various distributed power supply environments.

[0005] In order to solve the above technical problems, an embodiment of the present invention provides an optimization method for connecting distributed energy to a microgrid, including:

[0006] Input the current day's weather forecast data into the working model of each power generation equipment to obtain the output forecast data of each power generation equipment;

[0007] The load data of the day is clustered based on similarity and coordination based on the load level matching strategy to obtain the load level types and load shedding strategies of different levels in different time periods.

[0008] According to the output forecast data of each power generation equipment, the load level type in different time periods and the load removal strategy of different levels, the operation mode strategy is established, and according to the working model of each distributed unit, the operation mode strategy, the total cost model of microgrid operation and various constraints, the microgrid daily dispatch optimization model is established;

[0009] The microgrid daily dispatch optimization model is solved by improving the particle swarm algorithm, and the microgrid daily optimal dispatch plan is obtained. Distributed energy is connected to the microgrid according to the microgrid daily optimal dispatch plan.

[0010] In the implementation of the embodiment of the present invention, the meteorological forecast data of the day is passed through the working model of each power generation equipment to obtain the output forecast data of each power generation equipment, and the load data of the day is subjected to similar collaborative clustering based on the load level matching strategy. The "load similarity" collaborative clustering method is used to obtain the load level type in different time periods and the load removal strategy of different levels under island operation, and the operation mode strategy is established according to the output forecast data of each power generation equipment, the load level type in different time periods and the load removal strategy of different levels. Then, according to the working model of each distributed unit, the operation mode strategy, the total cost model of microgrid operation and various constraints, a microgrid daily scheduling optimization model based on load level matching is established, and an improved particle swarm algorithm with a "destruction and repair" mechanism is introduced to solve the microgrid optimization model, and a microgrid daily optimization scheduling plan is obtained, so as to optimize various distributed energy sources connected to the microgrid, effectively manage the output of each distributed power source in the microgrid, ensure the power supply reliability of the microgrid, effectively reduce the supply and demand deviation rate of the microgrid's electrical and thermal loads, improve the overall operation efficiency of the microgrid, and improve the stability of the microgrid operation under a variety of distributed power supply environments.

[0011] As a preferred solution, the load data of the day is clustered based on similarity and coordination based on the load level matching strategy to obtain the load level types and load shedding strategies of different levels in different time periods, specifically:

[0012] Several load levels are preset, and several independent power supply reliability classifications and several independent load interruption loss coefficient classifications corresponding to each load setting are set. A membership matrix is ​​established according to each power supply reliability classification and each load interruption loss coefficient classification;

[0013] The load similarity coefficient is introduced as a penalty coefficient to characterize the relationship strength between loads of different levels, and the improved value function is obtained according to the membership matrix and the load similarity coefficient. The formula is:

[0014]

[0015] Among them, U rde is the membership matrix corresponding to power supply reliability and load interruption loss coefficient, i and j are different types of load sets, i≠j, k is the type of class center, r is the type of load data point on a certain load set, M is the total number of class centers, N is the total number of load data points, z is the load level, and the z types of load levels are P[Ⅰ], P[Ⅱ], …, P[z], U rde-k,r [i] is the membership degree of the rth load data point on the i-th load set to the kth class, U rde-k,r [j] is the membership degree of the rth load data point on the jth load set to the kth class, x r [i] is the rth load data point in the i-th load set, y k [i] is the center of the kth type of load set, Xpsc [i,j] is the load similarity coefficient between different levels of loads;

[0016] The load data of the day is used as the input data of the clustering method, and the value of the improved value function is minimized by iterative operation of updating the cluster center and updating the membership degree, and the load level type of each time period is obtained; wherein, the load data of the day includes: the load demand of the day, the reliability of power supply and the load interruption loss coefficient;

[0017] Among them, the formula for updating the cluster center is:

[0018]

[0019] Update the membership, the formula is:

[0020]

[0021] in, k′ is the center position of the previous generation class, y k is the updated k-th category center, is the Euclidean distance between the center of the k′th type of load concentration and the data point of the rth type of load, is the Euclidean distance between the center of the kth type of load and the data point of the rth type of load after the i-th type of load concentration is updated;

[0022] According to the load level type in different time periods, the importance of the load in each time period is determined. When the power demand cannot be met under island operation, the loads of the corresponding level categories are removed in order from low to high according to the importance of the load in each time period to obtain load removal strategies of different levels.

[0023] As a preferred solution, the weather forecast data for the day is input into the working model of each power generation equipment to obtain the output forecast data of each power generation equipment, specifically:

[0024] The predicted wind speed for the day is input into the wind turbine output model to obtain the predicted wind turbine output data for the day; the wind turbine output model has the formula:

[0025]

[0026] Among them, P WT is the output power of the wind turbine unit, v is the predicted wind speed for the day, and v in is the minimum cut-in wind speed of the unit, v p is the rated wind speed of the unit, v out is the maximum cut-out wind speed of the unit, k 0 is the first working characteristic parameter of the fan unit, k 1 is the second working characteristic parameter of the fan unit, k 2 is the third working characteristic parameter of the fan unit, P tis the rated value of the fan unit output power;

[0027] The predicted sunlight and temperature for the day are input into the photovoltaic group output model to obtain the photovoltaic output forecast data for the day; the photovoltaic group output model has the formula:

[0028]

[0029] Among them, P PV is the output power of the photovoltaic unit, P STD is the maximum output power of the photovoltaic unit under standard conditions, lx in is the predicted light intensity for the day, lx STD is the standard light intensity, k is the power temperature coefficient, T c is the solar cell temperature, T r Forecast temperature for the day.

[0030] As a preferred solution, each distributed unit working model is a working model of each distributed unit in the microgrid; wherein each distributed unit includes power generation equipment, power supply equipment, heating equipment, energy storage device and refrigeration equipment. Except for the working model of each power generation equipment, the working model of the remaining distributed units is as follows:

[0031]

[0032] Among them, M work,i represents the working model of the i-th distributed unit, E sup represents the power supply equipment in the microgrid, Q sup represents the heating equipment in the microgrid, C sup represents the refrigeration unit in the microgrid, E device Represents the energy storage device in the microgrid, P MT is the output power of the gas turbine during operation, η MT is the total efficiency of the gas turbine, m f is the fuel flow rate, LHV is the lower heating value of natural gas, Q MT Recover heat power for gas turbine; η FC is the fuel cell efficiency, P FC Output active power for the battery; Q HRSG is the heat supply of the waste heat boiler, η HRSG is the heat conversion efficiency of the waste heat boiler; Q pv Output heat for photovoltaic thermal equipment, A s is the installation area of ​​photovoltaic collector equipment, η pv is the thermal conversion efficiency of photovoltaic collector equipment, SRI is the light radiation index; C c Provides cooling capacity model for refrigeration units, E c is the power consumption of the electric refrigeration unit, X ecis the refrigeration coefficient of the electric refrigeration unit, Q FC Generate heat for the fuel cell, η qc I is the refrigeration efficiency of the absorption refrigeration machine; e (t) is the actual charge and discharge current at time t, P bat (t) represents the total power provided by the energy storage device up to time t, U e represents the voltage value across the energy storage device, SOC(t) and SOC(t-1) are the state of charge values ​​of the energy storage device at time t and (t-1), η ch , η dis is the charging and discharging efficiency of the energy storage device up to time t, C N is the rated charge capacity of the energy storage device.

[0033] As a preferred solution, the total cost model of microgrid operation includes power supply cost model and environmental governance cost model, and the formula is:

[0034] f 3 =ω 1 F 1 +ω 2 F 2

[0035] Among them, F 1 The cost of supplying power to the microgrid, F 2 is the environmental governance cost of the microgrid, F 3 represents the total cost of microgrid operation, ω 1 is the first weighting coefficient, ω 2 is the second weighting coefficient, ω 1 ≥0,ω 2 ≥0, and ω 1 +ω 2 =1;

[0036] The power supply cost model is a multi-objective function constructed by combining the energy consumption cost of each distributed unit, maintenance cost, depreciation cost, power outage compensation cost and the interaction cost between the microgrid and the main grid. The formula is:

[0037]

[0038] Among them, F 1 The cost of supplying power to the microgrid, C 1 is the energy consumption cost of each distributed unit, C 2 is the maintenance cost of each distributed unit, C 3 is the depreciation cost of each distributed unit, C 4 Compensation for power outages for users, C 5 is the interaction cost between the microgrid and the main grid, T is the dispatching period, N is the number of power generation units in the microgrid, i is the type of distributed unit, C gasis the natural gas price, LHV is the lower heating value of natural gas, η i (t) represents the combustion efficiency of the i-th distributed unit at time t, X fix,i is the maintenance coefficient of distributed unit i, C ACC,i is the installation cost of distributed unit i, P n,i is the rated power of distributed unit i, f cf,i is the capacity factor of the i-th distributed unit, P i (t) represents the active power output by distributed unit i at time t, C L is the unit power failure compensation coefficient when the load is removed, P remove (t) is the load removal at time t, C 购电 (t) and C 售电 (t) represents the microgrid power purchase price and the main grid power sales price at time t, P 购-grid (t) and P 售-grid (t) is the amount of electricity purchased and sold by the microgrid from the main grid at time t, α is the microgrid consideration coefficient, when the microgrid is in grid-connected operation, there is energy exchange between the microgrid and the main grid, α=1, when the microgrid is in island operation, there is no energy exchange between the microgrid and the main grid, α=0, C grid Represents the transaction fee between the microgrid and the main grid;

[0039] The environmental governance cost model is a function constructed by integrating various environmental pollution control costs. The formula is:

[0040]

[0041] Among them, F 2 is the environmental governance cost of the microgrid, k is the type of pollutant, T is the scheduling period, N is the number of power generation units in the microgrid, i is the type of distributed unit, and M is the total number of pollutants; C k represents the cost of pollutant control, Z i,k is the penalty rate for generating pollutant k for the i-th distributed unit, P i (t) represents the active power output of distributed unit i at time t, a is the microgrid consideration coefficient, when the microgrid is in grid-connected operation, there is energy exchange between the microgrid and the main grid, a=1, when the microgrid is in island operation, there is no energy exchange between the microgrid and the main grid, a=0, Z grid,k is the penalty rate for pollutant k generated by the interaction between the microgrid and the main grid, P grid (t) is the interaction power between the microgrid and the main grid at time t.

[0042] As the preferred solution, various constraints include system energy balance constraint, distributed unit capacity constraint, microgrid and main grid interaction power constraint, capacity fluctuation constraint, pollutant emission constraint and energy storage unit power constraint. The formula is:

[0043]

[0044] in, is the j-level electric load in time period t, where j is level Ⅰ, level Ⅱ, and level Ⅲ. is the heat load of level j in period t, represents the output of each distributed unit in time period t, represents the interactive power between the microgrid and the main grid during the period t, T is the scheduling period, N is the number of power generation units in the microgrid, i is the type of distributed unit, is the minimum capacity of the i-th distributed unit, is the maximum capacity of the i-th distributed unit, is the minimum capacity for interaction between the microgrid and the main grid, is the maximum capacity of interaction between the microgrid and the main grid, P i max is the spare capacity of the i-th distributed unit during the scheduling period, P x,l (t) is the different load requirements at time t, L x % is the reserve capacity demand factor for different loads, P PV,WT (t) is the total output of photovoltaic and wind turbines at time t, u s % is the demand factor of photovoltaic and wind turbine for spare capacity, G i,k (P i ) is the value of various pollutants generated during the operation of each distributed unit, L k are the limiting values ​​of various pollutant emissions, is the output of the energy storage device at time t, is the minimum and maximum charging and discharging power of the energy storage device, and δ is the scheduling time period.

[0045] As a preferred solution, an operation mode strategy is established based on the output forecast data of each power generation equipment, the load level type in different time periods and the load removal strategy of different levels, specifically:

[0046] The operation mode strategy is selected according to the power situation in the microgrid, and the microgrid is separated from the main grid by disconnecting the static switch. The operation mode strategy is established according to the output forecast data of each power generation equipment, the load level type in different time periods and the load removal strategy of different levels; among which, the operation mode strategy includes the grid-connected operation mode strategy and the island operation mode strategy;

[0047] When running the grid-connected operation mode strategy, the microgrid is equivalent to the controllable load of the main grid. According to the power consumption on the load demand side, wind turbines and photovoltaic power supply are used first. When the power provided cannot meet the power demand, a scheduling plan is formulated to adjust the output of other units or purchase electricity from the main grid;

[0048] When running the island operation mode strategy, there is no power interaction between the microgrid and the main grid, and all electricity demand is provided by distributed units, with priority given to wind turbines and photovoltaic power supply. When the provided electricity cannot meet the electricity demand, low-cost power generation units will provide electricity. If it still cannot meet the electricity demand, load shedding will be carried out according to different levels of load shedding strategies.

[0049] As a preferred solution, the grid-connected operation mode strategy is as follows:

[0050]

[0051] Island operation mode strategy, specifically:

[0052]

[0053] Among them, T Ⅰ级负荷 Indicates the dispatching time period corresponding to level I load, T Ⅱ级负荷 Indicates the dispatching time period corresponding to the level II load, T Ⅲ级负荷 Indicates the dispatching time period corresponding to the Class III load, E PV,WT Indicates the power supply of photovoltaic and wind turbine, P el,Ⅰ , P ql,Ⅰ Indicates the level I electrical and thermal load demand, P el,Ⅱ , P ql,Ⅱ Indicates the electrical and thermal load demand of level II, P el,Ⅲ , P ql,Ⅲ Indicates the III-level electrical and thermal load requirements, price MT,FC represents the power supply cost of gas turbines and fuel cells, price BT Indicates the discharge cost of the battery, price 购电 represents the cost of purchasing electricity from the main grid, P MT,FC represents the power supply of the gas turbine and fuel cell, P BT Indicates the discharge capacity of the battery, P 购-grid Indicates the amount of electricity purchased from the main network, Q HRSG,pv Indicates the heating capacity of waste heat boiler and photovoltaic thermal collection equipment.

[0054] As a preferred solution, the daily dispatch optimization model of the microgrid is solved by improving the particle swarm algorithm to obtain the microgrid daily optimization dispatch plan, which is as follows:

[0055] The system energy balance constraint, distributed unit capacity constraint and energy storage unit power constraint are used as penalty factors to improve the total benefit function of the microgrid, and the objective function is improved as follows:

[0056]

[0057] Among them, λ is the penalty factor of the system energy balance constraint, σ is the penalty factor of the distributed unit capacity constraint and the energy storage unit power constraint;

[0058] The operation mode strategy gives priority to the use of wind turbine and photovoltaic output, initializes the input particles of the particle swarm algorithm with the daily output of the remaining distributed units, and optimizes the improved objective function according to the destruction operator and the repair operator to solve the microgrid daily scheduling optimization model and obtain the microgrid daily optimization scheduling plan; among them, the microgrid daily optimization scheduling plan includes the output data of each distributed unit, the system energy balance data and the comparison data of the total operation cost of the microgrid.

[0059] As a preferred solution, the destruction operator is formulated as follows:

[0060] R i,j =X eq,i (t)+V i,j +T i.j

[0061]

[0062] Among them, R i,j is the destruction coefficient of distributed energy solution i; X eq,i (t) is the load loss coefficient of distributed unit i at time t, including the electrical load loss coefficient and the thermal load loss coefficient; V i,j Is the supply match between distributed unit i and load level j, 1 if matched, 0 if not; T i.j P is whether the distributed unit i and load level j match in the scheduling phase, if they match, it is 1, if they do not match, it is 0; E,i (t) is the power supply of distributed unit i at time t; N E (t) is the electric load demand at time t; P Q,i (t) is the heat supply of distributed unit i at time t; N Q (t) is the heat load demand at time t;

[0063] Repair operator, the formula is:

[0064] z i =minΔF 3,i,j

[0065] Among them, z i represents the minimum repair cost, ΔF 3,i,j It represents the change in the total benefit of the microgrid caused by the repair of distributed unit i. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 : A flow chart of an embodiment of an optimization method for connecting distributed energy to a microgrid provided by the present invention;

[0067] Figure 2 : A schematic diagram of photovoltaic and wind turbine daily output curve prediction for an embodiment of an optimization method for connecting distributed energy to a microgrid provided by the present invention;

[0068] Figure 3 : A diagram of establishing a microgrid daily scheduling optimization model for an embodiment of an optimization method for connecting distributed energy to a microgrid provided by the present invention;

[0069] Figure 4 :A flowchart of solving a microgrid daily dispatch optimization model using an improved particle swarm algorithm according to an embodiment of an optimization method for connecting distributed energy to a microgrid provided by the present invention;

[0070] Figure 5 : An improved PSO algorithm for solving the output curve of the power supply distributed unit provided by the present invention in an embodiment of an optimization method for connecting distributed energy to a microgrid;

[0071] Figure 6 : An improved PSO algorithm for solving the output curve of the distributed heating unit provided by the present invention in an embodiment of an optimization method for connecting distributed energy to a microgrid;

[0072] Figure 7 :A system energy balance diagram based on a load level matching strategy for an embodiment of an optimization method for connecting distributed energy to a microgrid provided by the present invention;

[0073] Figure 8 : A comparison diagram of the total cost of microgrid operation under different strategies of an embodiment of an optimization method for connecting distributed energy to a microgrid provided by the present invention;

[0074] Fig. 9 : A schematic diagram of evaluation indicators of a microgrid daily scheduling optimization model of an embodiment of an optimization method for connecting distributed energy to a microgrid provided by the present invention. DETAILED DESCRIPTION

[0075] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0076] Embodiment 1

[0077] Please refer to Figure 1, which is a flow chart of an optimization method for connecting distributed energy to a microgrid provided by an embodiment of the present invention. The optimization method of this embodiment is applicable to distributed power supply of a microgrid. This embodiment optimizes the scheduling scheme of each distributed energy connected to the microgrid by establishing and solving a microgrid daily scheduling optimization model, thereby improving the stability and power supply reliability of the microgrid operation under a variety of distributed power supply environments. The optimization method includes steps 101 to 104, and each step is specifically as follows:

[0078] Step 101: input the weather forecast data for the day into the working model of each power generation equipment to obtain the output forecast data of each power generation equipment.

[0079] In this embodiment, the daily meteorological forecast data such as wind speed, light and temperature are obtained, and the working models of each power generation equipment are used. The working models of each power generation equipment include a wind turbine group output model and a photovoltaic group output model, to obtain the output forecast data of each power generation equipment (the output forecast data of the wind turbine and photovoltaic group on the day).

[0080] Optionally, step 101 specifically includes step 1011 to step 1012, and each step is specifically as follows:

[0081] Step 1011: input the predicted wind speed for the day into the wind turbine output model to obtain the predicted wind turbine output data for the day; wherein the wind turbine output model, i.e., the wind turbine output power model, is a formula of:

[0082]

[0083] Among them, P WT is the output power of the wind turbine unit, v is the predicted wind speed for the day, and v in is the minimum cut-in wind speed of the unit, v p is the rated wind speed of the unit, v out is the maximum cut-out wind speed of the unit, k 0 is the first working characteristic parameter of the fan unit, k 1 is the second working characteristic parameter of the fan unit, k 2 is the third working characteristic parameter of the fan unit, i.e. k 0 , k 1 and k 2 is the fan operating characteristic parameter, P t It is the rated value of the fan unit output power.

[0084] Step 1012: input the predicted sunlight and temperature for the day into the photovoltaic group output model to obtain the photovoltaic output prediction data for the day; wherein the photovoltaic group output model, that is, the photovoltaic group output power model, is formulated as follows:

[0085]

[0086] Among them, PPV is the output power of the photovoltaic unit (light intensity is lx in Output power at STD It is the maximum output power of the photovoltaic unit under standard conditions. Standard conditions generally refer to the light intensity lx STD =1000W / m 2 , temperature is 25℃, lx in is the predicted light intensity for the day, i.e., input light intensity, lx STD is the standard light intensity, k is the power temperature coefficient, T c is the solar cell temperature, T r It is the predicted temperature for the day, that is, the reference temperature.

[0087] In this embodiment, when the daily weather forecast such as wind speed, light, temperature, etc. is known, the output forecast data of each power generation equipment is obtained according to the wind turbine group output model and the photovoltaic group output model, and the wind turbine and photovoltaic output forecast curves of the day are obtained by linear interpolation. The photovoltaic and wind turbine daily output curve forecast schematic diagram is as follows: Figure 2 As shown, the output of the wind turbine fluctuates greatly in different time periods, the photovoltaic unit is in the output state from 6 to 19 o'clock, the wind turbine output is relatively large from 0 to 5 o'clock, and the photovoltaic unit output is relatively large from 10 to 15 o'clock.

[0088] Step 102: Perform similar collaborative clustering on the load data of the day based on the load level matching strategy to obtain the load level types and load shedding strategies of different levels in different time periods.

[0089] In this embodiment, the load data of the day is obtained, including the load demand of the day, power supply reliability and load interruption loss coefficient (resilience weight index), and the "load similarity" collaborative clustering method is used to obtain the load level type in different time periods and the removal strategy of different levels of load under island operation, that is, the load demand of the day P l-e,q , power supply reliabilityδ Rel and the load interruption loss factor δ R-loss As the input data of the clustering method, the improved value function C(U rde ,i,j)′ value is the smallest, and the load level type in different time periods (the time period corresponding to each type of load level) can be obtained. At the same time, the importance of the load in each time period is determined according to the obtained indicators. When the island is running, if the power demand cannot be met, the load of the corresponding level category is removed in order from low to high according to the load importance level, and different levels of load removal strategies are obtained.

[0090] Optionally, step 102 specifically includes step 1021 to step 1024, each step is as follows:

[0091] Step 1021: preset several load levels, and set several independent power supply reliability classifications and several independent load interruption loss coefficient classifications corresponding to each load level. According to each power supply reliability classification and each load interruption loss coefficient classification, establish a membership matrix.

[0092] In this embodiment, there are z load levels preset as P[I], P[II], ..., P[z], and each load has an independent power supply reliability classification. and separate load interruption loss factor classification The corresponding membership matrix is

[0093] Step 1022: Introduce the load similarity coefficient as a penalty coefficient to characterize the relationship strength between loads of different levels, and obtain an improved value function based on the membership matrix and the load similarity coefficient. The formula is:

[0094]

[0095] Among them, U rde Power supply reliability Rel and load interruption loss factor δ R-loss The corresponding membership matrix, i, j are different types of load sets, i≠j, k is the type of class center, r is the type of load data points on a certain load set, M is the total number of class centers, N is the total number of load data points, z is the load level, and the z types of load levels are P[Ⅰ], P[Ⅱ], ..., P[z], U rde-k,r [i] is the membership degree of the rth load data point on the i-th load set to the kth class, U rde-k,r [j] is the membership degree of the rth load data point on the jth load set to the kth class, x r [i] is the rth load data point in the i-th load set, y k [i] is the center of the kth type of load set, X psc [i,j] is the load similarity coefficient between different levels of loads.

[0096] Step 1023: using the load data of the day as the input data of the clustering method, and minimizing the value of the improved value function by iteratively updating the cluster center and the membership degree, and obtaining the load level type by time period; wherein the load data of the day includes: the load demand of the day, the power supply reliability and the load interruption loss coefficient;

[0097] Among them, the formula for updating the cluster center is:

[0098]

[0099] Update the membership, the formula is:

[0100]

[0101] in, k′ is the center position of the previous generation class, y k is the updated k-th category center, is the Euclidean distance between the center of the k′th type of load concentration and the data point of the rth type of load, It is the Euclidean distance between the center of the kth type of load and the data point of the rth type of load after the i-th type of load is concentrated and updated.

[0102] Step 1024: Determine the importance of the load in each time period according to the load level type in each time period. When the power demand cannot be met under island operation, cut off the loads of the corresponding level categories in order from low to high according to the importance of the load in each time period to obtain load cutting strategies of different levels.

[0103] In this embodiment, the optimal clustering effect is obtained through iteration, which is k=3. The corresponding 24-hour electricity and heat load demands are divided into level I, level II, and level III. Level I indicates the highest load demand and importance, and level III indicates the lowest load demand and importance. When the load demand cannot be met under island operation, level III load should be removed first.

[0104] As an example of this embodiment, the time interval for each load demand statistics is 1 hour, and 24 hours are divided into 24 / 1=24 time periods. The "load similarity" collaborative clustering method is used to obtain the load level type and load removal range of different time periods, as shown in Table 1.

[0105] Table 1 Load level type and load removal range in different time periods

[0106]

[0107] According to Table 1, the type of Class I load is high electric load demand and high heat load demand, and the corresponding time periods are 12-15 and 18-21; the type of Class II load is medium electric load demand and medium heat load demand, and the corresponding time periods are 8-11, 16-17, and 22; the type of Class III load is low electric load demand and low heat load demand, and the corresponding time periods are 1-7 and 23-24; in the island operation mode, when each unit in the microgrid still cannot meet the load power and heat demand when running at full load, it will be removed according to the power supply reliability and load interruption loss coefficient. The range of Class I electric load removal is 0-3.2kw, and the range of heat load removal is 0-1.4kw. The range of Class II electric load removal is 0-7.5kw, and the range of heat load removal is 0-2.9kw. The range of Class III electric load removal is 0-10.1kw, and the range of heat load removal is 0-5.2kw.

[0108] Step 103: Establish an operation mode strategy based on the output forecast data of each power generation equipment, the load level type in different time periods and the load removal strategy of different levels, and establish a microgrid daily scheduling optimization model based on the working model of each distributed unit, the operation mode strategy, the microgrid operation total cost model and various constraints.

[0109] In this embodiment, a microgrid daily dispatch optimization model is established, such as Figure 3 As shown in the figure, according to the output forecast data of each power generation equipment (photovoltaic and wind turbine output forecast), the load level type in different time periods and the load removal strategies of different levels, the operation mode strategy is established. According to the working model of each distributed unit, the operation mode strategy, the total cost model of microgrid operation and various constraints, a microgrid optimization model based on load level matching is established, and a microgrid dispatching strategy, namely the operation mode strategy, is formulated. The microgrid dispatching strategy includes the grid-connected operation mode level dispatching strategy (grid-connected operation mode strategy) and the island operation mode level dispatching strategy (island operation mode strategy).

[0110] Optionally, each distributed unit working model is a working model of each distributed unit in the microgrid; wherein each distributed unit includes power generation equipment, power supply equipment, heating equipment, energy storage device and refrigeration equipment, photovoltaic and wind turbines belong to power generation equipment, and their working models have been listed in step 1011 and step 1012, and will not be repeated here. In addition to the working models of each power generation equipment, the working models of the remaining distributed units are as follows:

[0111]

[0112] Among them, M work,i represents the working model of the i-th distributed unit, E sup represents the power supply equipment in the microgrid, Q sup represents the heating equipment in the microgrid, C sup represents the refrigeration unit in the microgrid, E device Represents the energy storage device in the microgrid, P MT is the output power of the gas turbine during operation, η MT is the total efficiency of the gas turbine, m f is the fuel flow rate, LHV is the lower heating value of natural gas, Q MT Recover heat power for gas turbine; η FC is the fuel cell efficiency, P FC Output active power for the battery; Q HRSG is the heat supply of the waste heat boiler, η HRSG is the heat conversion efficiency of the waste heat boiler; Q pv Output heat for photovoltaic thermal equipment, A s is the installation area of ​​photovoltaic collector equipment, η pvis the thermal conversion efficiency of photovoltaic collector equipment, SRI is the light radiation index; C c Provides cooling capacity model for refrigeration units, E c is the power consumption of the electric refrigeration unit, X ec is the refrigeration coefficient of the electric refrigeration unit, Q FC Generate heat for the fuel cell, η qc I is the refrigeration efficiency of the absorption refrigeration machine; e (t) is the actual charge and discharge current at time t, P bat (t) represents the total power provided by the energy storage device up to time t, U e represents the voltage value across the energy storage device, SOC(t) and SOC(t-1) are the state of charge values ​​of the energy storage device at time t and (t-1), η ch , η dis is the charging and discharging efficiency of the energy storage device up to time t, C N is the rated charge capacity of the energy storage device.

[0113] Optionally, the total cost model of microgrid operation includes a power supply cost model and an environmental governance cost model, and the formula is:

[0114] f 3 =ω 1 F 1 +ω 2 F 2

[0115] Among them, F 1 The cost of supplying power to the microgrid, F 2 is the environmental governance cost of the microgrid, F 3 represents the total cost of microgrid operation, the total cost of microgrid operation, ω 1 is the first weighting coefficient, ω 2 is the second weighting coefficient, ω 1 ≥0,ω 2 ≥0, and ω 1 +ω 2 =1, considering that power supply cost is as important as environmental governance, ω 1 =ω 2 =1 / 2.

[0116] The power supply cost model is a multi-objective function constructed by combining the energy consumption cost of each distributed unit, maintenance cost, depreciation cost, power outage compensation cost and the interaction cost between the microgrid and the main grid. The formula is:

[0117]

[0118] Among them, F 1 The cost of supplying power to the microgrid, C 1 is the energy consumption cost of each distributed unit, C2 is the maintenance cost of each distributed unit, C 3 is the depreciation cost of each distributed unit, C 4 Compensation for power outages for users, C 5 is the interaction cost between the microgrid and the main grid, T is the dispatching period, N is the number of power generation units in the microgrid, i is the type of distributed unit, C gas is the natural gas price, LHV is the lower heating value of natural gas, η i (t) represents the combustion efficiency of the i-th distributed unit at time t, X fix,i is the maintenance coefficient of distributed unit i, C ACC,i is the installation cost of distributed unit i, P n,i is the rated power of distributed unit i, f cf,i is the capacity factor of the i-th distributed unit, P i (t) represents the active power output by distributed unit i at time t, C L is the unit power failure compensation coefficient when the load is removed, P remove (t) is the load removal at time t, C 购电 (t) and C 售电 (t) represents the microgrid power purchase price and the main grid power sales price at time t, P 购-grid (t) and P 售-grid (t) is the amount of electricity purchased and sold by the microgrid from the main grid at time t, α is the microgrid consideration coefficient, when the microgrid is in grid-connected operation, there is energy exchange between the microgrid and the main grid, α=1, when the microgrid is in island operation, there is no energy exchange between the microgrid and the main grid, α=0, C grid Represents the transaction fee between the microgrid and the main grid.

[0119] The environmental governance cost model is a function constructed by integrating various environmental pollution control costs. The formula is:

[0120]

[0121] Among them, F 2 is the environmental governance cost of the microgrid, k is the type of pollutant, T is the scheduling period, N is the number of power generation units in the microgrid, i is the type of distributed unit, and M is the total number of pollutants; C k represents the cost of pollutant control, Z i,k is the penalty rate for generating pollutant k for the i-th distributed unit, P i (t) represents the active power output of distributed unit i at time t, a is the microgrid consideration coefficient, when the microgrid is in grid-connected operation, there is energy exchange between the microgrid and the main grid, a=1, when the microgrid is in island operation, there is no energy exchange between the microgrid and the main grid, a=0, Z grid,kis the penalty rate for pollutant k generated by the interaction between the microgrid and the main grid, P grid (t) is the interaction power between the microgrid and the main grid at time t.

[0122] Optionally, various constraints include system energy balance constraints, distributed unit capacity constraints, microgrid and main grid interaction power constraints, capacity fluctuation constraints, pollutant emission constraints and energy storage unit power constraints. The formula is:

[0123]

[0124] in, is the j-level electric load in time period t, where j is level Ⅰ, level Ⅱ, and level Ⅲ. is the heat load of level j in period t, represents the output of each distributed unit in time period t, represents the interactive power between the microgrid and the main grid during the period t, T is the scheduling period, N is the number of power generation units in the microgrid, i is the type of distributed unit, is the minimum capacity of the i-th distributed unit, is the maximum capacity of the i-th distributed unit, is the minimum capacity for interaction between the microgrid and the main grid, is the maximum capacity of interaction between the microgrid and the main grid, is the spare capacity of the i-th distributed unit during the scheduling period, P x,l (t) is the different load demands at time t, L x % is the reserve capacity demand factor for different loads, P PV,WT (t) is the total output of photovoltaic and wind turbines at time t, u s % is the demand factor of photovoltaic and wind turbine for spare capacity, G i,k (P i ) is the value of various pollutants generated during the operation of each distributed unit, L k are the limiting values ​​of various pollutant emissions, is the output of the energy storage device at time t, is the minimum and maximum charging and discharging power of the energy storage device, and δ is the scheduling time period.

[0125] Optionally, an operation mode strategy is established based on the output forecast data of each power generation equipment, the load level type in different time periods, and the load shedding strategies of different levels, specifically:

[0126] The operation mode strategy is selected according to the power situation in the microgrid, and the microgrid is separated from the main grid by disconnecting the static switch. The operation mode strategy is established according to the output forecast data of each power generation equipment, the load level type in different time periods and the load removal strategy of different levels; among which, the operation mode strategy includes the grid-connected operation mode strategy and the island operation mode strategy;

[0127] When running the grid-connected operation mode strategy, the microgrid is equivalent to the controllable load of the main grid. According to the power consumption on the load demand side, wind turbines and photovoltaic power supply are used first. When the power provided cannot meet the power demand, a scheduling plan is formulated to adjust the output of other units or purchase electricity from the main grid;

[0128] When running the island operation mode strategy, there is no power interaction between the microgrid and the main grid, and all electricity demand is provided by distributed units, with priority given to wind turbines and photovoltaic power supply. When the provided electricity cannot meet the electricity demand, low-cost power generation units will provide electricity. If it still cannot meet the electricity demand, load shedding will be carried out according to different levels of load shedding strategies.

[0129] In this embodiment, the selection of the grid-connected or islanded operation mode of the microgrid is mainly based on the power situation (such as: abnormal voltage value in the grid, power quality, etc.), and the microgrid is separated from the main grid by disconnecting the static switch. During grid-connected operation, the microgrid is equivalent to the controllable load of the main grid. According to the power consumption on the load demand side, wind turbines and photovoltaic power supply are given priority. When the power provided cannot meet the power demand, a scheduling plan is formulated to adjust the output of other units or purchase electricity from the main grid, thereby reducing the total cost of microgrid operation. Under island operation, there is no power interaction between the microgrid and the main grid, and all power demand is provided by distributed units. Wind turbines and photovoltaic power supply are also given priority. When the power provided cannot meet the power demand, the power is provided by units with lower power generation costs. If it still cannot meet the power demand, the load is removed according to different levels of load removal strategies, thereby ensuring the reliability of microgrid power supply.

[0130] Optional, grid-connected operation mode strategy, specifically:

[0131]

[0132] Island operation mode strategy, specifically:

[0133]

[0134] Among them, T Ⅰ级负荷 Indicates the dispatching time period corresponding to level I load, T Ⅱ级负荷 Indicates the dispatching time period corresponding to the level II load, T Ⅲ级负荷 Indicates the dispatching time period corresponding to the Class III load, E PV,WT Indicates the power supply of photovoltaic and wind turbine, P el,Ⅰ , P ql,Ⅰ Indicates the level I electrical and thermal load demand, P el,Ⅱ , P ql,Ⅱ Indicates the electrical and thermal load demand of level II, P el,Ⅲ , P ql,Ⅲ Indicates the III-level electrical and thermal load requirements, price MT,FCrepresents the power supply cost of gas turbines and fuel cells, price BT Indicates the discharge cost of the battery, price 购电 represents the cost of purchasing electricity from the main grid, P MT,FC represents the power supply of the gas turbine and fuel cell, P BT Indicates the discharge capacity of the battery, P 购-grid Indicates the amount of electricity purchased from the main network, Q HRSG,pv Indicates the heating capacity of waste heat boiler and photovoltaic thermal collection equipment.

[0135] In this embodiment, when t∈T Ⅰ级负荷 When the grid-connected operation is carried out according to P MT,FC →P BT →P 购-grid Compensation for Level I electricity demand P el,Ⅰ , the island runs according to min{price MT,FC ,price BT} corresponding to P MT,FC or P BT Compensation for Level I electricity demand P el,Ⅰ , Q HRSG,pv Matching the heat demand of level I; when t∈T Ⅱ级负荷 When the grid-connected operation is carried out according to P BT →P 购-grid →P MT,FC Compensation for Level II electricity demand P el,Ⅱ , island operation with t∈T Ⅰ级负荷 The phase is the same, but the remaining power needs to be considered for P BT Charging, Q HRSG Matching the heat demand of level II; when t∈T Ⅲ级负荷 When connected to the grid, PV and WT are used to the maximum extent, and the surplus power is used to charge BT and the main grid P 售-grid The island operation strategy is consistent with the grid-connected operation strategy, but it is impossible to sell electricity to the main grid at this time. pv Matching the heat demand of level III.

[0136] Step 104: Solve the microgrid daily dispatch optimization model by improving the particle swarm algorithm to obtain the microgrid daily optimization dispatch plan, and connect the distributed energy to the microgrid according to the microgrid daily optimization dispatch plan.

[0137] Optionally, the daily dispatch optimization model of the microgrid is solved by improving the particle swarm algorithm to obtain the daily optimal dispatch plan of the microgrid, which is specifically:

[0138] The system energy balance constraint, distributed unit capacity constraint and energy storage unit power constraint (unit ramp constraint) are used as penalty factors to improve the total benefit function of the microgrid, and the objective function is improved as follows:

[0139]

[0140] Among them, F 3 is the total benefit function of the microgrid system (total cost of microgrid operation), λ is the penalty factor of the system energy balance constraint, and σ is the penalty factor of the distributed unit capacity constraint and the energy storage unit power constraint;

[0141] The operation mode strategy gives priority to the use of wind turbine and photovoltaic output, initializes the input particles of the particle swarm algorithm with the daily output of the remaining distributed units, and optimizes the improved objective function according to the destruction operator and the repair operator to solve the microgrid daily scheduling optimization model and obtain the microgrid daily optimization scheduling plan; among them, the microgrid daily optimization scheduling plan includes the output data of each distributed unit, the system energy balance data and the comparison data of the total operation cost of the microgrid.

[0142] In this embodiment, the flowchart of the improved particle swarm algorithm for solving the microgrid daily dispatch optimization model is as follows: Figure 4 As shown in the figure, the microgrid optimization model solved by the improved particle swarm algorithm with the idea of ​​"destruction and repair" is introduced. Wind turbine and photovoltaic output are given priority in both grid-connected and islanded operation. The daily output conditions of the remaining distributed units are initialized and input into the particles. The objective function is constructed by using the load classification strategy (load level type in different time periods and load removal strategy of different levels) and the scheduling strategy (operation mode strategy). The objective function is optimized and solved to obtain the daily optimal scheduling plan of the microgrid, including the output data of each distributed unit, the system energy balance data and the comparison data of the total operation cost of the microgrid, that is, the output curve of each distributed unit, the system energy balance diagram and the comparison diagram of the total operation cost of the microgrid.

[0143] Optional, destruction operator, the formula is:

[0144] R i,j =X eq,i (t)+V i,j +T i.j

[0145]

[0146] Among them, R i,j is the destruction coefficient of distributed energy solution i. The higher the destruction coefficient, the greater the destruction probability of the solution. eq,i (t) is the load loss coefficient of distributed unit i at time t, including the electrical load loss coefficient and the thermal load loss coefficient; V i,j Is the supply match between distributed unit i and load level j, 1 if matched, 0 if not; T i.j P is whether the distributed unit i and load level j match in the scheduling phase, if they match, it is 1, if they do not match, it is 0; E,i (t) is the power supply of distributed unit i at time t; NE (t) is the electric load demand at time t; P Q,i (t) is the heat supply of distributed unit i at time t; N Q (t) is the heat load demand at time t;

[0147] Repair operator, the formula is:

[0148] z i =minΔF 3,i,j

[0149] Among them, z i represents the minimum repair cost, ΔF 3,i,j It represents the change in the total benefit of the microgrid caused by the repair of distributed unit i.

[0150] In this embodiment, an improved particle swarm algorithm of "destruction and repair" is introduced to solve the microgrid daily dispatch optimization model, that is, the improved objective function is optimized according to the destruction operator and the repair operator to solve the microgrid daily dispatch optimization model as follows:

[0151]

[0152] Among them, x i (P MT,BT,Grid,..., Q pv,HRSG ) represents the daily output electric power and thermal power of each distributed unit solved by the improved particle swarm PSO algorithm, P i,s.t , Q i,s,i represents the electric power output point and thermal power output point removed from a distributed unit at a certain time obtained according to the destruction operator, x i-destroy Indicates removal of P i,s.t , Q i,s,i The solution set after I 1 ,I 2 Indicates P i,s.t and Q i,s,i The damage coefficient is sorted from low to high, z i represents the minimum repair cost, pos i,t P represents the removal i,s.t and Q i,s,i In the original solution set x i The position at the moment, x i-repair Represents the solution set after the "repair" operation.

[0153] In this embodiment, the daily output data of the distributed unit is obtained by improving the particle swarm PSO algorithm, and the damage operator R is used to calculate the daily output data of the distributed unit. i,jThe calculation formula is used to solve the destruction coefficient of each distributed unit, and the number of destruction solutions is randomly determined. The unreasonable output points are taken out in order from large to small according to the destruction coefficient and their scheduling times are recorded. The daily output data of the "destroyed" distributed unit is empty at certain times. Then, the concentrated output data points of the destruction solution are repaired, that is, they are reasonably inserted into the original empty positions to minimize the repair cost, and finally the daily output data of each distributed unit after further optimization is obtained.

[0154] For example, the improved PSO algorithm is used to solve the output curve of the distributed power supply unit, such as Figure 5 As shown in the figure, Figures S / R1 to S / R6 respectively represent the locations where the random electric load "destruction and repair" operator is generated. It can be seen from the output curve of the power supply distributed unit that in the level I electric load scheduling stage, the output of photovoltaic and wind turbines cannot meet the electricity demand. At this time, gas turbines and fuel cells are used to supplement the power supply; in the level III electric load scheduling stage, the output of photovoltaic and wind turbines is used to the maximum extent, and the remaining power is used to charge the energy storage device and sell electricity to the main grid; at 16-24, the microgrid is working in an isolated operation state and has no interaction with the main grid; the trend of the output curve of the power supply distributed unit intuitively shows the scheduling strategy of different levels of load.

[0155] Figure 6 To improve the PSO algorithm to solve the output curve of the distributed heating unit, S / Rq1~S / Rq4 in the figure respectively represent the positions where the random heat load "destruction and repair" operator is generated. The photovoltaic collector equipment is in a non-output state from 19:00 to 24:00, and this stage is the peak time for heat use, so the output burden of the waste heat boiler is increased. The output curve trend of the heating unit intuitively shows the matching relationship of the corresponding level of load.

[0156] Figure 7 This is a system energy balance diagram based on the load level matching strategy, which shows the balance relationship between the system electricity and heat load demand and supply in each period. When the microgrid is in grid-connected operation mode, it can balance energy through the purchase and sale of electricity from the main grid. In the level I electricity load dispatching stage, it mainly sells electricity to the main grid, and in the level III electricity load dispatching stage, it mainly purchases electricity from the main grid; when the microgrid is in island operation mode, there is no connection with the main grid. If the power supply and heat supply cannot meet the demand at this time, it is necessary to unload a certain amount of electricity and heat load to maintain the system energy balance; the system energy balance directly reflects the economy and stability of the microgrid operation. If the system works in a high output state for a long time, it will cause a large amount of electricity and heat load loss, and affect the service life of the unit; if the system works in a low output state for a long time, user demand cannot be met, affecting the stability of the system.

[0157] Figure 8This is a comparison chart of the total cost of microgrid operation under different strategies. In the figure, S / R1~S / R6 respectively represent the locations where the random electric load "destruction repair" operator is generated, and S / Rq1~S / Rq4 respectively represent the locations where the random thermal load "destruction repair" operator is generated. As can be seen from the figure, in the level I load scheduling stage, the total cost of microgrid operation of the patent scheduling strategy (operation mode strategy) is higher than that of the conventional strategy at some moments, while in the level II and level III load scheduling stages, the total cost of microgrid operation of the patent scheduling strategy is lower than that of the conventional strategy. The total operating costs of the conventional strategy at different levels of load are 845.67 yuan, 766.72 yuan, and 1022.59 yuan respectively, and the total operating costs of the patent scheduling strategy at different levels of load are 835.13 yuan, 722.80 yuan, and 951.47 yuan respectively. The total cost of distributed energy access to the microgrid is increased through the patent scheduling strategy, saving operating costs.

[0158] In this embodiment, the dispatch optimization scheme can be evaluated. According to the distributed unit output curve, the system energy balance diagram and the microgrid operation total cost comparison diagram, the dispatch optimization scheme is evaluated. The power and heat load supply and demand deviation index and the economic benefit improvement rate are used as the data source for the long-term dispatch optimization of the microgrid dispatch model, and the optimization effect of the microgrid daily dispatch optimization model is further improved. The power and heat load supply and demand deviation index φ match-i,j and economic benefit improvement rate As the data source for the long-term iterative training of the subsequent scheduling model, the specific steps are as follows:

[0159]

[0160] Among them, x i-repair represents the solution set after the “repair” operation, fig(p i,T ) represents the output curve of each distributed unit, fig(System balance ) represents the system energy balance diagram, fig(F 3-contrast ) represents the comparison of the total operation cost of the microgrid with the dispatch strategy and the conventional strategy. Indicates the deviation rate between the power supply and heating units and the electric load and heat load demand in different scheduling periods. It represents the economic benefit improvement rate of the dispatching strategy and the conventional dispatching strategy at different load levels, OPT long (F,i,j,T long ) represents the long-term dispatch optimization plan of the microgrid.

[0161] By improving the particle swarm algorithm, the output data of each distributed unit in the final microgrid is obtained, the corresponding data points are plotted as images, and the deviation rates of power supply and heating units and electric load and thermal load demand in different scheduling periods as well as the economic benefit improvement rates in different load stages are calculated. These index data can be used as the basis for long-term scheduling of microgrids, that is, as the basis for improving or adjusting quarterly optimization and annual optimization models.

[0162] For example, Fig. 9 This is a schematic diagram of the evaluation indicators of the microgrid daily scheduling optimization model. When the supply-demand deviation rate is a positive value, it means that the supply of electricity and heat loads is greater than the demand. When the supply-demand deviation rate is a negative value, it means that the supply of electricity and heat loads is less than the demand. When the economic benefit improvement rate is a positive value, it means that the total operating cost of the microgrid of this patent strategy is lower than that of the conventional strategy. When the economic benefit improvement rate is a negative value, it means that the total operating cost of the microgrid of this patent strategy is higher than that of the conventional strategy. Due to the load level classification, the supply-demand deviation rate and the economic benefit improvement rate of this patent strategy are significantly improved compared with the conventional strategy.

[0163] Calculate the evaluation index at different load levels, and the deviation rate between the power supply and heating units of the microgrid and the electric load and heat load demand at different dispatching periods Economic benefit improvement rate of dispatching strategy and conventional dispatching strategy at different load levels The comparison data of evaluation indicators is shown in Table 2 below.

[0164] Table 2 Comparison data of evaluation indicators

[0165]

[0166] It can be seen from Table 2 that the supply and demand deviation rates of the conventional strategy under different levels of electric load are 6.98%, 5.26% and 4.29% respectively, the supply and demand deviation rates of the patent strategy under different levels of electric load are 4.95%, 2.89% and 1.41% respectively, the supply and demand deviation rates of the conventional strategy under different levels of thermal load are 10.26%, 8.58% and 5.67% respectively, the supply and demand deviation rates of the patent strategy under different levels of thermal load are 9.11%, 6.07% and 4.03% respectively. Compared with the conventional strategy, the supply and demand deviation rates of the electric load in different levels of load stages of this patent are reduced by 2.03%, 2.37% and 2.88% respectively, and the supply and demand deviation rates of the thermal load are reduced by 1.15%, 2.51% and 1.64% respectively. The total benefits of microgrid operation are increased by 1.25%, 5.73% and 6.95% respectively, and the improvement effect is significant in the level II and level III load scheduling stages.

[0167] In implementing the embodiment of the present invention, first, the "load similarity" collaborative clustering method is used to obtain the load level type in different time periods and the load removal strategy of different levels under island operation. Then, a microgrid daily scheduling optimization model based on load level matching is established, and the particle swarm algorithm improved by the "destruction repair" mechanism is introduced to solve the microgrid daily scheduling optimization model, and then the evaluation index is used as the data source for long-term scheduling optimization of the microgrid daily scheduling optimization model, so as to further improve the stability of the microgrid. Finally, a microgrid simulation example is built in MATLAB. The simulation results show that this optimization method can effectively reduce the supply and demand deviation rate of the microgrid's electric and thermal loads and improve the overall operation efficiency of the microgrid. Compared with the conventional scheduling optimization strategy, the supply and demand deviation rate of the electric load in different load stages is reduced by 2.03%, 2.37% and 2.88% respectively, and the supply and demand deviation rate of the thermal load is reduced by 1.15%, 2.51% and 1.64% respectively. The total operation efficiency of the microgrid is improved by 1.25%, 5.73% and 6.95% respectively, which verifies the effectiveness and practicality of this method. Optimize the various distributed energy sources connected to the microgrid, effectively manage the output of various distributed power sources within the microgrid, ensure the power supply reliability of the microgrid, effectively reduce the supply and demand deviation rate of the microgrid's electrical and thermal loads, improve the overall operating efficiency of the microgrid, and improve the stability of the microgrid's operation under a variety of distributed power supply environments.

[0168] The specific embodiments described above further illustrate the purpose, technical solutions and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of protection of the present invention.

Claims

1. An optimization method for connecting distributed energy to microgrids. It is characterized in that include: Input the current day's weather forecast data into the working model of each power generation equipment to obtain the output forecast data of each power generation equipment; The load data of the day is clustered based on similarity and coordination based on the load level matching strategy to obtain the load level types and load shedding strategies of different levels in different time periods. Establish an operation mode strategy according to the output forecast data of each power generation equipment, the load level type in each time period and the load removal strategy of different levels, and establish a microgrid daily scheduling optimization model according to the working model of each distributed unit, the operation mode strategy, the total cost model of microgrid operation and various constraints; Solving the microgrid daily dispatch optimization model by improving the particle swarm algorithm to obtain a microgrid daily optimization dispatch plan, and connecting distributed energy to the microgrid according to the microgrid daily optimization dispatch plan; The operation mode strategy is established according to the output forecast data of each power generation equipment, the time-divided load level type and the load shedding strategy of different levels, specifically: The operation mode strategy is selected according to the power situation in the microgrid, and the microgrid is separated from the main grid by disconnecting the static switch, and the operation mode strategy is established according to the output forecast data of each power generation equipment, the time-divided load level type and the load removal strategy of different levels; wherein the operation mode strategy includes a grid-connected operation mode strategy and an island operation mode strategy; When the grid-connected operation mode strategy is run, the microgrid is equivalent to the controllable load of the main grid. According to the power consumption on the load demand side, wind turbines and photovoltaic power supply are used preferentially. When the power provided cannot meet the power demand, a scheduling plan is formulated to adjust the output of other units or purchase electricity from the main grid; When the island operation mode strategy is running, the microgrid has no power interaction with the main grid, and all electricity demands are provided by distributed units. The wind turbines and photovoltaic power supply are used preferentially. When the provided electricity cannot meet the electricity demand, the low-cost power generation units provide electricity. If the electricity demand still cannot be met, the load is shelved according to the different levels of load shedding strategies.

2. The optimization method for connecting distributed energy to a microgrid as claimed in claim 1, It is characterized in that The load data of the day is subjected to similar collaborative clustering based on the load level matching strategy to obtain the load level type and load shedding strategies of different levels in different time periods, specifically: Preset several load levels, set several independent power supply reliability classifications and several independent load interruption loss coefficient classifications corresponding to each load, and establish a membership matrix according to each power supply reliability classification and each load interruption loss coefficient classification; The load similarity coefficient is introduced as a penalty coefficient to characterize the relationship strength between loads of different levels, and the improved value function is obtained according to the membership matrix and the load similarity coefficient. The formula is: in, is the membership matrix corresponding to power supply reliability and load interruption loss coefficient, , For different load set types, , is the class center type, is the type of load data point on a load set, is the total number of cluster centers, is the total number of load data points, is the load level, The load level is , , …, , For the Load set The load data point for The class membership, For the Load set The load data point for The class membership, For the Load set load data points, For the Load set Type Center, is the load similarity coefficient between different levels of loads; The load data of the day is used as the input data of the clustering method, and the value of the improved value function is minimized by iterative operation of updating the cluster center and updating the membership degree, so as to obtain the load level type of the time period; wherein the load data of the day includes: the load demand of the day, the power supply reliability and the load interruption loss coefficient; The formula for updating the cluster center is: The updated membership is formulated as follows: in, , , The center position of the previous generation class, After the update Type Center, For the The load concentration Type center and The Euclidean distance of the load data points, For the After the load is concentrated and updated Type center and Euclidean distance of load data points; According to the load level type in each time period, the importance of the load in each time period is determined. When the power demand cannot be met under island operation, the loads of corresponding levels are cut off in order from low to high according to the importance of the load in each time period to obtain the load cutting strategies of different levels.

3. The optimization method for connecting distributed energy to a microgrid as claimed in claim 1, It is characterized in that The weather forecast data for the day is input into the working model of each power generation equipment to obtain the output forecast data of each power generation equipment, specifically: The predicted wind speed for the day is input into the wind turbine output model to obtain the predicted wind turbine output data for the day; wherein the wind turbine output model has the formula: in, is the output power of the fan unit, is the predicted wind speed for the day, is the minimum cut-in wind speed of the unit, is the rated wind speed of the unit, is the maximum cut-out wind speed of the unit, is the first working characteristic parameter of the fan unit, is the second working characteristic parameter of the fan unit, is the third working characteristic parameter of the fan unit, is the rated value of the fan unit output power; The predicted sunlight and temperature for the day are input into the photovoltaic group output model to obtain the photovoltaic output forecast data for the day; wherein the photovoltaic group output model has the formula: in, is the output power of the photovoltaic unit, is the maximum output power of the photovoltaic unit under standard conditions, For the predicted sunlight for the day, is the standard light intensity, is the power temperature coefficient, is the solar cell temperature, The predicted temperature for the day.

4. The optimization method for connecting distributed energy to a microgrid as claimed in claim 1, It is characterized in that The distributed unit working models are working models of the distributed units in the microgrid; wherein the distributed units include power generation equipment, power supply equipment, heating equipment, energy storage devices and refrigeration equipment. Except for the working models of the power generation equipment, the working models of the remaining distributed units are as follows: in, Indicates A distributed unit working model, Represents the power supply equipment in the microgrid, represents the heating equipment in the microgrid, represents the refrigeration unit in the microgrid, represents the energy storage device in the microgrid, is the output power of the gas turbine during operation, is the total efficiency of the gas turbine, is the fuel flow rate, The low calorific value of natural gas. Recover thermal power for gas turbines; For fuel cell efficiency, Output active power to the battery; The heat supply of the waste heat boiler is is the heat conversion efficiency of the waste heat boiler; Output heat for photovoltaic thermal equipment, is the installation area of ​​photovoltaic collector equipment, is the thermal conversion efficiency of photovoltaic collector equipment, is the light radiation index; Provides cooling capacity model for refrigeration units, The electricity consumed by the electric refrigeration unit, is the refrigeration coefficient of the electric refrigeration unit, Generates heat for the fuel cell, The refrigeration efficiency of the absorption refrigeration machine; for t The actual charge and discharge current at each moment, Indicates that the deadline t The energy storage device provides total power at all times. Indicates the voltage value across the energy storage device. , For energy storage devices t Time and ( t -1) the state of charge value at the moment, , For the deadline t The charging and discharging efficiency of the energy storage device at all times, C N is the rated charge capacity of the energy storage device.

5. The optimization method for connecting distributed energy to a microgrid as claimed in claim 1, It is characterized in that The total cost model of microgrid operation includes a power supply cost model and an environmental governance cost model, and the formula is: in, The cost of powering a microgrid, The environmental governance cost of microgrids is represents the total cost of microgrid operation, is the first weighting coefficient, is the second weighting coefficient, ≥0, ≥0, and ; The power supply cost model is a multi-objective function constructed by combining the energy consumption cost of each distributed unit, maintenance cost, depreciation cost, power outage compensation cost and the interaction cost between the microgrid and the main grid. The formula is: in, The cost of supplying power to the microgrid, is the energy consumption cost of each distributed unit, is the maintenance cost of each distributed unit, is the depreciation cost of each distributed unit, Compensation for power outages for users, is the interaction fee between the microgrid and the main grid, is the scheduling period, is the number of power generation units in the microgrid, For the distributed unit type, is the natural gas price, Indicates the lower calorific value of natural gas. Indicates Distributed units in t Combustion efficiency at all times, Distributed Unit The maintenance factor, Distributed Unit The installation cost, Distributed Unit Rated power, For the The capacity factor of the distributed unit, Represents a distributed unit exist t The active power output at any time, is the unit power outage compensation coefficient when the load is cut off, for t Cut off the load at all times. and express t The microgrid power purchase price and the main grid power sales price at the moment, and for t The amount of electricity purchased and sold by the microgrid from the main grid at any given moment. Consider the coefficient of the microgrid. When the microgrid is connected to the grid, there is energy exchange between the microgrid and the main grid. =1, when the microgrid is operating in an isolated island, there is no energy exchange between the microgrid and the main grid. =0, Represents the transaction fee between the microgrid and the main grid; The environmental governance cost model is a function constructed by integrating various environmental pollution control costs, and the formula is: in, The environmental governance cost of microgrids is is the type of pollutant, is the scheduling period, is the number of power generation units in the microgrid, For the distributed unit type, is the total number of pollutants; represents the cost of pollutant control, For the Distributed units generate pollutants The penalty rate, Represents a distributed unit exist t The active power output at any time, Consider the coefficient of the microgrid. When the microgrid is connected to the grid, there is energy exchange between the microgrid and the main grid. =1, when the microgrid is operating in an isolated island, there is no energy exchange between the microgrid and the main grid. =0, Pollutants generated by the interaction between microgrid and main grid The penalty rate, for t The microgrid interacts with the main grid at all times.

6. The optimization method for connecting distributed energy to a microgrid as claimed in claim 1, It is characterized in that The various constraints include system energy balance constraints, distributed unit capacity constraints, microgrid and main grid interaction power constraints, capacity fluctuation constraints, pollutant emission constraints and energy storage unit power constraints. The formula is: in, for t During the period Level electrical load, Take level Ⅰ, level Ⅱ, level Ⅲ, for t During the period Heat load, express t The output of each distributed unit during the period, express t The interactive power between the microgrid and the main grid during the period, is the scheduling period, is the number of power generation units in the microgrid, For the distributed unit type, For the The minimum capacity of the distributed unit, For the The maximum capacity of the distributed unit, is the minimum capacity for interaction between the microgrid and the main grid, is the maximum capacity of interaction between the microgrid and the main grid, For the The spare capacity of the distributed units during the dispatch period, for t Different load demands at all times, is the reserve capacity requirement factor for different loads, For photovoltaic and wind turbines t Total effort at all times, is the demand factor of photovoltaic and wind turbine for spare capacity, are the values ​​of various pollutants generated during the operation of each distributed unit. are the limiting values ​​of various pollutant emissions, for t The energy storage device is outputting at all times. , is the minimum and maximum charging and discharging power of the energy storage device, The scheduling time period.

7. The optimization method for connecting distributed energy to a microgrid as claimed in claim 1, It is characterized in that The grid-connected operation mode strategy is specifically as follows: The island operation mode strategy is specifically as follows: in, Indicates the dispatching time period corresponding to level I load, Indicates the dispatching time period corresponding to the level II load, Indicates the dispatching time period corresponding to the level III load, Indicates the power supply of photovoltaic and wind turbines, , Indicates the level I electrical and thermal load demand, , Indicates the II-level electrical and thermal load requirements, , Indicates the III-level electrical and thermal load requirements. represents the power supply cost of gas turbines and fuel cells, represents the discharge cost of the battery, represents the cost of purchasing electricity from the main grid, Indicates the power supply of the gas turbine and fuel cell, Indicates the discharge capacity of the battery. Indicates the amount of electricity purchased from the main network. Indicates the heating capacity of waste heat boiler and photovoltaic thermal collection equipment.

8. The optimization method for connecting distributed energy to a microgrid as claimed in claim 6, It is characterized in that The microgrid daily dispatch optimization model is solved by improving the particle swarm algorithm to obtain the microgrid daily optimization dispatch plan, which is specifically: The system energy balance constraint, the distributed unit capacity constraint and the energy storage unit power constraint are used as penalty factors to improve the total benefit function of the microgrid, and the objective function is improved, and the expression is: in, is the penalty factor for the energy balance constraint of the system, is a penalty factor for the distributed unit capacity constraint and the energy storage unit power constraint; The operation mode strategy gives priority to the use of wind turbine and photovoltaic output, initializes the input particles of the particle swarm algorithm with the daily output of the remaining distributed units, and optimizes the improved objective function according to the destruction operator and the repair operator to solve the microgrid daily scheduling optimization model to obtain a microgrid daily optimization scheduling plan; wherein, the microgrid daily optimization scheduling plan includes the output data of each distributed unit, the system energy balance data and the comparison data of the total cost of microgrid operation.

9. The optimization method for connecting distributed energy to a microgrid as claimed in claim 8, It is characterized in that The damage operator is formulated as follows: in, Distributed Energy The damage coefficient of the solution; for t Time Distributed Unit The load loss coefficient includes the electrical load loss coefficient and the thermal load loss coefficient; Distributed Unit With load level Whether the supply matches, 1 if it matches, 0 if it does not match; Distributed Unit With load level Whether the scheduling phase matches, if it matches, it is 1, if it does not match, it is 0; for t Time Distributed Unit The amount of power supply; for t The electricity load demand at each moment; is the distributed unit at time t of heating supply; for t Heat load demand at any given moment; The repair operator is formulated as follows: in, represents the minimum repair cost, Represents a distributed unit The change in the total benefit of the microgrid after repair.

Images