Microgrid optimization scheduling method based on lexicographical order multi-target prediction control

Abstract

The invention relates to a micro-grid optimization scheduling method based on lexicographical order multi-target prediction control, and the method comprises the steps: constructing a grid-connected micro-grid system connected with a main power grid, and carrying out the modeling of components in the system; establishing a multi-objective function of micro-grid operation, and determining a lexicographical order priority order according to the relative importance of the objectives; measuring values of the state of the micro-grid system are continuously obtained within scheduling time, optimization problem solving is carried out through a lexicographical order MO-MPC algorithm, an optimal control sequence is obtained, and a first element of the optimal control sequence is applied to the system. According to the invention, multiple objectives of power balance, economic cost and environmental protection are realized, and energy production of the micro-grid is effectively managed while economic benefits and environmental benefits are considered, so that the load demand at any moment is met; and the selection of the multifarious multi-objective function priority weight coefficient is avoided, so that the processing of the priority multi-objective optimization problem becomes simple.

Description

Technical Field

[0001] The present invention relates to the technical field of calculation, extrapolation or counting, and in particular to a microgrid optimization scheduling method based on lexicographic order multi-objective predictive control. Background Art

[0002] With the development of society, the demand for electricity continues to increase. Considering the environmental and economic problems caused by the use of fossil energy, renewable energy as an alternative solution to energy demand has attracted more and more attention.

[0003] Renewable energy can be recycled in nature. Common renewable energy sources include wind energy and solar energy, which have been widely used in energy supply. However, renewable energy sources such as wind energy and solar energy are often distributed, random, and intermittent, and cannot provide continuous and stable power output. One way to solve this problem is to combine these intermittent energy sources with energy storage equipment, other forms of distributed generation, and local power loads into a microgrid subsystem to stabilize power output and increase the absorption of renewable energy.

[0004] Micro-Grid refers to a small power generation and distribution system composed of distributed power sources, energy storage devices, energy conversion devices, loads, monitoring and protection devices, etc. It aims to achieve flexible and efficient application of distributed power sources and solve the problem of connecting a large number of distributed power sources in various forms to the grid. Micro-grids can operate independently of the main grid or in parallel with the main grid. During the process of connecting to the grid, micro-grids can also buy or sell energy from the main grid.

[0005] In addition, in order to maximize the use of renewable energy for power generation and meet electricity demand, it is necessary to optimize the power scheduling of microgrids. As an important part of smart grid optimization, it is of great significance to reduce energy consumption and environmental pollution.

[0006] However, in actual operation, microgrid optimization scheduling is a complex multi-objective optimization problem. Simply considering economic costs, environmental benefits or the tracking performance of a certain indicator cannot fully describe the operation of the microgrid, and some goals are conflicting. How to coordinate these multi-objective problems according to priority and obtain the overall optimal solution is one of the key issues that need to be solved in microgrid optimization scheduling. For the multi-objective function optimization problem in microgrid optimization scheduling, existing methods usually weight different objective functions when dealing with its priority problem, and characterize the target priority through weight coefficients. However, the selection of this set of weight coefficients is relatively difficult and requires a large number of offline experiments. In addition, in the presence of constraints, the size of the weight coefficient may not necessarily make the obtained solution accurately meet the priority of the target.

[0007] How to better coordinate the conflicting multi-objective optimization problems in microgrid optimal scheduling is a key issue that needs to be solved at present. Summary of the invention

[0008] The present invention solves the problems existing in the prior art and provides a microgrid optimization scheduling method based on lexicographically ordered multi-objective predictive control (MO-MPC), that is, the multi-objective functions are sorted in lexicographic order according to the priority, and the solutions are optimized sequentially, and the obtained solutions meet the priorities of the multi-objective functions and avoid the selection of weight coefficients; model predictive control has the advantages of online processing of optimization problems, rolling optimization, and explicit processing of multiple constraints, and can optimize the economic indicators of the system. It is one of the main control methods for managing the stable economic operation of microgrids.

[0009] The technical solution adopted by the present invention is a microgrid optimization scheduling method based on lexicographic multi-objective predictive control, which constructs a grid-connected microgrid system connected to the main power grid, and models the components in the system respectively; establishes a multi-objective function for characterizing the operation of the system, and determines the lexicographic priority ranking according to the relative importance of the objectives;

[0010] The measured values ​​of the microgrid system state are continuously obtained within the dispatch time, and the optimization problem is solved by the lexicographic MO-MPC (Multi-objective model predictive control) algorithm to obtain the optimal control sequence, and the first element of the optimal control sequence is applied to the system.

[0011] Preferably, the microgrid system includes a wind turbine, a photovoltaic generator, a micro gas turbine and an energy storage battery, and a load is provided in conjunction with the microgrid system.

[0012] Preferably, component modeling of the microgrid system includes:

[0013] Based on wind power generation power P w The mathematical model of wind turbine generator established based on wind speed v and photovoltaic power generation power P pv and solar irradiance S and ambient temperature T air The mathematical model of photovoltaic generator, the dynamic model of energy storage battery and the mathematical model of micro-turbine fuel cost were established.

[0014] Preferably, based on historical data and a mathematical model of a wind turbine generator and a mathematical model of a photovoltaic generator, the wind power generation power and the photovoltaic power generation power are predicted;

[0015] Taking the state of charge SOC of the energy storage battery as the state quantity, x(k)=[SOC(k)];

[0016] The main power grid power P grid , Micro-turbine power PMT And the energy storage battery power P bat is the control quantity, u(k)=[P grid (k), P MT (k), P bat (k)] T , T represents vector transpose, and the state space model is established.

[0017] x(k+1)=Ax(k)+Bu(k)

[0018] y(k)=x(k)

[0019] Among them, A and B are coefficient matrices, k is the time; x(k) is x k , u(k) is u k .

[0020] Preferably, the multi-objective function includes a tracking objective function J1, an economic objective function J2 and an environmental objective function J3;

[0021] The multi-objective functions are prioritized in lexicographic order according to their relative importance.

[0022] Preferably, the tracking objective function is:

[0023]

[0024] Among them, P L (k) = P pv (k)+P w (k)+P MT (k)+P grid (k)+P bat (k), P pv (k) is the photovoltaic power generation power at time k, P w (k) is the wind power generation power at time k; u k,N =[u 0|k ,u 1|k ,…u N-1|k ] T ,u i|k represents the predicted value at time k for time k+i, N represents the prediction time domain, and a=[1 1 1].

[0025] Preferably, the economic objective function is:

[0026]

[0027] Wherein, b = [α1, α2, 0], α1 and α2 are the price coefficients of the main power grid and the micro-turbine respectively.

[0028] Preferably, the environmental objective function is:

[0029]

[0030] Among them, α, β, γ are the emission coefficients of the micro-turbine, c = [0 1 0], R = diag (0, 1, 0).

[0031] Preferably, setting constraints includes:

[0032] SOC min ≤SOC(k)≤SOC max

[0033] u min ≤u k ≤u max

[0034] △u min ≤△u k ≤△u max

[0035] Among them, SOC(k) is the state of charge of the energy storage battery at time k, SOC min is the minimum power level, SOC max is the maximum power level; u min To provide the minimum power, u max is the maximum value of the power supply; Δu min and Δu max The upper and lower limits of power change.

[0036] Preferably, the single-objective optimization problem is solved in sequence according to the lexicographic priority to obtain the overall optimal control sequence. According to the rolling optimization principle, the first element of the solved lexicographic optimal sequence is applied to the system, and the above solving steps are repeated to achieve rolling optimization.

[0037] The present invention relates to a microgrid optimization scheduling method based on lexicographic multi-objective predictive control, which comprises the following steps: constructing a grid-connected microgrid system connected to a main power grid, modeling components in the system respectively, establishing a multi-objective function for the operation of the microgrid, and determining a lexicographic priority ranking according to the relative importance of the objectives, continuously obtaining the measured values ​​of the microgrid system state within the scheduling time, solving the optimization problem with a lexicographic MO-MPC algorithm, obtaining an optimal control sequence, and applying the first element of the optimal control sequence to the system.

[0038] The beneficial effect of the present invention is that, for different tracking, economic and environmental targets in the operation of a microgrid, a lexicographic priority sorting is performed according to the relative importance of the targets, and a lexicographic MO-MPC method is adopted to express the optimization problem as a lexicographic finite time domain optimal control problem, which is solved sequentially through the hierarchical structure of a single-objective finite time domain optimal control problem to obtain a Parato optimal solution that satisfies the priorities of the tracking, economic and environmental control targets, and the microgrid power is optimally dispatched with the optimal solution to achieve multiple targets of power balance, economic cost and environmental protection, and the energy production of the microgrid is effectively managed while taking into account economic and environmental benefits to meet the load demand at any time; the present invention avoids the selection of complicated priority weight coefficients of multi-objective functions, making the processing of priority multi-objective optimization problems simple. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a schematic diagram of the microgrid structure of the present invention;

[0040] Figure 2 is a flow chart of the method of the present invention;

[0041] Figure 3 It is a flow chart of the lexicographic order multi-objective predictive control algorithm in the present invention;

[0042] Figure 4 It is a schematic diagram of the power output of each unit within the dispatching time, as well as the power generation of new energy and the power demand of load;

[0043] Figure 5 It is a schematic diagram of the battery charge state change during the scheduling time;

[0044] Figure 6 Schematic diagram of power generation tracking load demand during the scheduling time. DETAILED DESCRIPTION

[0045] The present invention is further described in detail below in conjunction with embodiments, but the protection scope of the present invention is not limited thereto.

[0046] The present invention relates to a microgrid optimization dispatching method based on lexicographic multi-objective predictive control, which constructs a grid-connected microgrid system connected to a main power grid, models the components in the system respectively, establishes a multi-objective function for characterizing the operation of the system, and determines the lexicographic priority order according to the relative importance of the objectives;

[0047] The measured values ​​of the microgrid system state are continuously obtained within the dispatch time, and the optimization problem is solved by the lexicographic MO-MPC algorithm to obtain the optimal control sequence, and the first element of the optimal control sequence is applied to the system.

[0048] In the present invention, each component in the system is modeled. The system consists of wind power, photovoltaic power, traditional micro-gas turbine, energy storage battery and local load, such as Figure 1 As shown in the figure, the renewable energy power generation is predicted based on the wind speed and solar irradiance of historical weather. The energy storage battery SOC (State of charge, SOC) is used as the state quantity, and the main grid power P grid , Micro-turbine power P MT and battery power P bat A state space model is established for the control quantity; corresponding objective functions and constraints are established according to the tracking problems, economic problems and environmental problems of the microgrid system operation; priorities are sorted according to the relative importance of different objectives without setting priority weight coefficients, and a lexicographic MO-MPC (Multi-objective model predictive control) method that can explicitly consider the priorities of different objectives is adopted. The MO-MPC problem is expressed as a lexicographic finite-time domain optimal control problem, which is solved through the hierarchical structure of the single-objective finite-time domain optimal control problem to obtain the lexicographic optimal control sequence of the entire problem. The solution automatically satisfies the lexicographic priority and the first element of the optimal sequence is applied to the system.

[0049] The microgrid system includes a wind turbine, a photovoltaic generator, a micro gas turbine and an energy storage battery, and a load is provided in conjunction with the microgrid system. It should be noted that k in the implementation mode represents the current time;

[0050] The component modeling of the microgrid system includes wind power generation power P w The mathematical model of wind turbine generator established based on wind speed v and photovoltaic power generation power P pv and solar irradiance S and ambient temperature T air The mathematical model of photovoltaic generator, the dynamic model of energy storage battery and the mathematical model of micro-turbine fuel cost were established.

[0051] Predict wind power generation and photovoltaic power generation based on historical data and mathematical models of wind turbines and photovoltaic generators;

[0052] Taking the state of charge SOC of the energy storage battery as the state quantity, x(k)=[SOC(k)];

[0053] The main power grid power P grid , Micro-turbine power P MT And the energy storage battery power P bat is the control quantity, u(k)=[P grid (k), P MT (k), P bat (k)] T, T represents vector transpose, and the state space model is established;

[0054] x(k+1)=Ax(k)+Bu(k)

[0055] y(k)=x(k)

[0056] Among them, A and B are coefficient matrices, k is the time; x(k) is x k , u(k) is u k . .

[0057] (1-1) According to the principle of wind power generation, the wind power generation power P is established w and mathematical model of wind speed v;

[0058]

[0059] Among them, P w (k) is the wind power generation power, v(k) is the wind speed at time k, v i is the cut-in wind speed, v o is the cut-out wind speed, v r is the rated wind speed, P r is the rated power;

[0060] (1-2) According to the principle of photovoltaic power generation, establish the photovoltaic power generation power P pv And solar irradiance S, ambient temperature T air Mathematical model of

[0061]

[0062] Among them, P pv (k) is the photovoltaic power generation, S(k) is the solar irradiance, T air (k) is the ambient temperature, P STC is the maximum test power under standard test conditions, S STC is the solar irradiance under standard test conditions, k1 is the power temperature coefficient, k2 is the solar irradiance coefficient. The coefficient setting here is based on actual needs. STC is the reference temperature under standard test conditions;

[0063] (1-3) Establish a dynamic model of the SOC of the energy storage battery;

[0064]

[0065] Among them, SOC represents the battery state of charge, η is the battery charge and discharge coefficient, usually 0.8 to 0.9, T s is the sampling time, which is determined according to the actual discrete step length, C max is the maximum battery capacity, P bat(t) is the battery power; when the battery supplies power to the microgrid, P bat (t) is positive, when the power is absorbed from the microgrid, P bat (t) is negative;

[0066] (1-4) Establish a mathematical model for micro-turbine fuel cost;

[0067]

[0068] Among them, C CH4 is the unit natural gas price, P MT (k) is the output power of the micro-turbine, T s is the sampling time, η MT is the efficiency of the micro-turbine, Q LHV It is the lower calorific value of natural gas, usually 0.8 or 0.85.

[0069] (1-5) Predicting wind speed and sunshine in the future based on historical wind speed and sunshine. The prediction of renewable power generation here can be achieved through the existing LSTM neural network prediction technology;

[0070] (1-6) With the battery SOC as the state variable x(k) = [SOC(k)], the main grid power P grid , Micro-turbine power P MT and battery power P bat is the control quantity, u(k)=[P grid (k), P MT (k), P bat (k)] T , T represents vector transpose, and the state space model is established

[0071]

[0072] Where A=1,B=[0 0-ηT s / C max ]; let x(k) be x k , u(k) is u k .

[0073] In the present invention, a state space model is established as a prediction model according to the operation of each component in the microgrid, and control is subsequently performed based on the prediction model.

[0074] In the actual operation of a microgrid, it is necessary to optimize the operation of the microgrid according to some performance indicators or economic indicators to obtain the best operating state and economic performance, and the operation constraints of the microgrid need to be satisfied during the optimization process; the multi-objective function includes a tracking objective function J1, an economic objective function J2, and an environmental objective function J3; according to the relative importance of the objectives, the multi-objective function is sorted in lexicographic order of priority. For example, in the general implementation process, pri(J3) < pri(J2) < pri(J1), that is, the tracking objective has a higher priority than the economic objective, and the economic objective has a higher priority than the environmental objective.

[0075] (2-1) In a microgrid, it is necessary to ensure that the output power can track the load demand and satisfy

[0076] P L (k) = P pv (k) + P w (k) + P MT (k) + P grid (k) + P bat (k) (6)

[0077] At time k, P pv (k) is the photovoltaic power generation, P w (k) is the wind power generation, P MT (k) is the output power of the micro gas turbine, P grid is the main grid power, P bat (k) is the power provided by the energy storage battery at the current time;

[0078] Based on this, a tracking objective function J1 is established. According to the power interaction with the main grid and the output of the micro gas turbine, an economic cost function J2 is established. According to the fuel cost during the operation of the micro gas turbine, an environmental objective function J3 is established;

[0079] The tracking objective function is

[0080]

[0081] where u k,N = [u 0|k , u 1|k , … u N-1|k T , u i|k represents the predicted value from time k to time k + i, N represents the prediction horizon, and a = [1 1 1];

[0082] The main economic cost in the microgrid comes from the power provided by the main grid and the micro gas turbine. The economic objective function is

[0083]

[0084] ​Wherein, b = [α1, α2, 0], α1 and α2 are the price coefficients of the main grid and the micro-turbine, respectively, which generally refer to the electricity price and natural gas price between the local microgrid and the main grid, such as 0.7 and 2.5 respectively;

[0085] The environmental impact caused by the microgrid mainly comes from the fuel combustion emissions of the micro-turbine. The environmental objective function is:

[0086]

[0087] Among them, α, β, γ are the emission coefficients of the micro-turbine. The values ​​of α, β, γ are obtained by data fitting based on the fuel consumption curve or data table provided by the micro-turbine manufacturer actually used, or by referring to the fuel consumption model and parameters of typical micro-turbines provided by existing literature, standard data sets or research materials, and selecting close values; c = [0 1 0], R = diag (0, 1, 0).

[0088] (2-2) Establishing SOC constraints according to the optimal level range of the battery SOC;

[0089] In order to reduce the loss of the battery and extend its service life, the SOC level of the battery is usually limited to the optimal operating range, that is,

[0090] SOC min ≤SOC(k)≤SOC max (10)

[0091] Among them, SOC min is the minimum power level, SOC max is the maximum power level;

[0092] Because of the physical limitations of the equipment, the power that can be provided is also limited. According to the power limitations of the micro-turbines, batteries, and the main power grid, power constraints are established;

[0093] u min ≤u k ≤u max (11)

[0094] In order to protect the equipment, the power change value should not be too large. According to the power change limit, establish the power change constraint

[0095] △u min ≤△u k ≤△u max (12)

[0096] Among them, Δu k =u k -u k-1 ;u minTo provide the minimum power, u max is the maximum value of the power supply; Δu min and Δu max The upper and lower limits of power change.

[0097] The single-objective optimization problem is solved in sequence according to the lexicographic priority to obtain the overall optimal control sequence. According to the rolling optimization principle, the first element of the solved lexicographic optimal sequence is applied to the system, and the above solving steps are repeated to achieve rolling optimization.

[0098] In the present invention, in order to solve tracking problems, economic profit problems, and environmental problems with different relative importance levels, a Parato optimal solution that meets multiple control target priorities is determined, and a lexicographic order MO-MPC method that can explicitly consider different target priorities is adopted, so that the solution automatically meets the target priorities without setting the weight coefficient of the objective function priority.

[0099] (3-1) Set the prediction time domain N, and prioritize the tracking problem target J1, economic profit target J2, and environmental problem target J3 according to the actual operation of the system;

[0100] In the actual operation of the microgrid system, first of all, power balance must be ensured, that is, the power provided by the system must accurately track the local load; secondly, the operation of the system must ensure the lowest economic cost; finally, under the premise of ensuring the above goals, the pollution caused by micro-turbine emissions must be minimized as much as possible; therefore, the lexicographic priority of the objective function is sorted as follows:

[0101] J=[J1(u k,N ), J2(u k,N ), J3(u k,N )] (13)

[0102] (3-2) For the set of objective functions with determined priorities, the lexicographic MO-MPC algorithm is used to solve the problem and obtain the lexicographic optimal control sequence of the entire optimization problem.

[0103] The MO-MPC problem is formulated as a lexicographically ordered finite-horizon optimal control problem, which is solved through a hierarchy of single-objective finite-horizon optimal control problems.

[0104] Solve the highest priority tracking problem first

[0105]

[0106] The optimal value obtained is used as one of the constraints of the next priority target optimization problem to solve the next priority economic target

[0107]

[0108] Add the optimal value of each priority obtained above to the constraint conditions of the next priority target to solve the next priority environmental target

[0109]

[0110] Get the optimal control sequence for the entire problem

[0111]

[0112] Among them, ε i (i=1,2) indicates the tolerance introduced for the convenience of calculation, which can be taken as ε i =0.05J i * ;

[0113] (3-3) The optimal control sequence The first element of is applied to the system,

[0114]

[0115] Then repeat the above steps to scroll the optimization run on the timeline.

[0116] The application logic of the present invention is to first establish the model of each component of the microgrid, and predict the power generation of renewable energy according to the historical wind speed and light; then establish the state space model of the system; establish the objective function and constraints according to the tracking performance, economic profit and environmental problems in the operation of the microgrid; prioritize the objective function; set the prediction time domain N, total time T, sampling time T s , set the initial time k = 0; measure the current state, solve the optimization problem in sequence, and obtain the optimal control sequence of the entire problem The first element of the lexicographically optimal control sequence Apply to the system, set k=k+1, return and repeat the above steps.

[0117] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A microgrid optimization scheduling method based on lexicographic multi-objective predictive control, characterized by: Build a grid-connected microgrid system connected to the main grid and model the components in the system separately; Establish a multi-objective function to characterize the system operation and determine the lexicographic priority order according to the relative importance of the objectives; The measured values ​​of the microgrid system state are continuously obtained within the dispatch time, and the optimization problem is solved by the lexicographic MO-MPC algorithm to obtain the optimal control sequence, and the first element of the optimal control sequence is applied to the system.

2. A microgrid optimization scheduling method based on lexicographic multi-objective predictive control according to claim 1, characterized in that: The microgrid system includes a wind turbine, a photovoltaic generator, a micro gas turbine and an energy storage battery, and a load is provided in conjunction with the microgrid system.

3. The microgrid optimization scheduling method based on lexicographic multi-objective predictive control according to claim 2 is characterized by: The component modeling of the microgrid system includes wind power generation power P w The mathematical model of wind turbine generator established based on wind speed v and photovoltaic power generation power P pv and solar irradiance S and ambient temperature T air The mathematical model of photovoltaic generator, the dynamic model of energy storage battery and the mathematical model of micro-turbine fuel cost were established.

4. The microgrid optimization scheduling method based on lexicographic multi-objective predictive control according to claim 2 is characterized by: Predict wind power generation and photovoltaic power generation based on historical data and mathematical models of wind turbines and photovoltaic generators; Taking the state of charge SOC of the energy storage battery as the state quantity, x(k)=[SOC(k)]; The main power grid power P grid , Micro-turbine power P MT And the energy storage battery power P bat is the control quantity, u(k)=[P grid (k), P MT (k), P bat (k)] T , T represents vector transpose, and the state space model is established. x(k+1)=Ax(k)+Bu(k) y(k)=x(k) Among them, A and B are coefficient matrices, k is the time; x(k) is x k , u(k) is u k .

5. The microgrid optimization scheduling method based on lexicographic multi-objective predictive control according to claim 1 is characterized by: The multi-objective function includes a tracking objective function J1, an economic objective function J2 and an environmental objective function J3; The multi-objective functions are prioritized in lexicographic order according to their relative importance.

6. A microgrid optimization scheduling method based on lexicographic multi-objective predictive control according to claim 5, characterized in that: The tracking objective function is: Among them, P L (k) = P pv (k)+P w (k)+P MT (k)+P grid (k)+P bat (k), P pv (k) is the photovoltaic power generation power at time k, P w (k) is the wind power generation power at time k; u k,N =[u 0|k ,u 1|k ,…u N-1|k ] T ,u i|k represents the predicted value at time k for time k+i, N represents the prediction time domain, and a=[1 1 1].

7. The microgrid optimization scheduling method based on lexicographic multi-objective predictive control according to claim 5 is characterized by: The economic objective function is: Wherein, b = [α1, α2, 0], α1 and α2 are the price coefficients of the main power grid and the micro-turbine respectively.

8. The microgrid optimization scheduling method based on lexicographic multi-objective predictive control according to claim 5 is characterized by: The environmental objective function is: Among them, α, β, γ are the emission coefficients of the micro-turbine, c = [0 1 0], R = diag (0, 1, 0).

9. The microgrid optimization scheduling method based on lexicographic multi-objective predictive control according to claim 5 is characterized by: Set constraints, including: SOC min ≤SOC(s)≤SOC max in min in k in max △u min ≤△u k ≤△u max Among them, SOC(k) is the state of charge of the energy storage battery at time k, SOC min is the minimum power level, SOC max is the maximum power level; u min To provide the minimum power, u max is the maximum value of the power supply; Δu min and Δu max The upper and lower limits of power change.

10. The microgrid optimization scheduling method based on lexicographic multi-objective predictive control according to claim 5 is characterized by: The single-objective optimization problem is solved in sequence according to the lexicographic priority to obtain the overall optimal control sequence. According to the rolling optimization principle, the first element of the solved lexicographic optimal sequence is applied to the system, and the above solving steps are repeated to achieve rolling optimization.

Images