Micro-grid optimization scheduling method and device based on multi-target particle swarm optimization algorithm

Abstract

The invention provides a micro-grid optimization scheduling method and device based on a multi-objective particle swarm optimization algorithm. The method comprises the following steps: taking predicted values of a renewable micro source and a load in a micro-grid in 24 hours of the next day and output limit values of energy storage equipment and a diesel engine as an optimization space of decision variables of an objective function of operation cost and environment cost, and initializing a particle swarm in the optimization space; obtaining a non-dominated solution according to the fitness value of the initialized particle swarm, and establishing an external archive REP; searching for an optimal particle leader in the gbest from the REP according to a roulette strategy; updating the inertia factor and the variation rate according to the current iteration frequency, executing PSO particle swarm operation, obtaining the updated next generation particle position and speed, updating the pbest, and updating the external archive REP; when the number of iterations reaches the maximum number of iterations, outputting an external archive REP to obtain a Pareto optimal solution set; and querying a compromise solution with the minimum operation cost and the minimum environment cost in the Pareto optimal solution set to serve as a next-day scheduling strategy scheme of the micro-grid.

Description

Technical Field

[0001] The present disclosure relates to the technical field of microgrid optimization scheduling, and in particular to a microgrid optimization scheduling method and device based on a multi-objective particle swarm optimization algorithm. Background Art

[0002] With the transformation of the global energy structure and the rapid development of renewable energy technology, microgrids, as a new type of power system integrating distributed generation, energy storage, and load management, play an increasingly important role in energy supply. However, the non-controllable renewable energy (such as photovoltaics, wind power, etc.) in microgrids has significant volatility and uncertainty, which brings great challenges to the optimization and scheduling of microgrids.

[0003] Traditional microgrid optimization scheduling methods usually adopt a single optimization objective, such as minimizing operating costs or environmental costs, and are often based on a longer scheduling cycle (such as hours or days). However, this scheduling method has the following problems when dealing with the volatility and uncertainty of non-controllable renewable energy: first, the power prediction accuracy under long periods is low, because climate and environmental changes over long time scales are more complex and difficult to predict; second, long-term scheduling is difficult to adapt to the rapidly changing load demand and renewable energy output in microgrids.

[0004] Therefore, developing a multi-objective optimization scheduling method that can accurately predict the power generation and load power of non-controllable renewable energy in the short term in the future and comprehensively consider the operating cost and environmental cost is of great significance for improving the operating efficiency of microgrids, reducing operating costs and reducing environmental pollution. Summary of the invention

[0005] The purpose of the present invention is to overcome the shortcomings of the prior art and provide a microgrid optimization scheduling method and device based on a multi-objective particle swarm optimization algorithm. The method can effectively solve the problems existing in traditional microgrid optimization scheduling methods, improve the operating efficiency and environmental protection performance of the microgrid, and has important practical application value.

[0006] According to a first aspect of an embodiment of the present disclosure, a microgrid optimization scheduling method based on a multi-objective particle swarm optimization algorithm is provided, the method comprising:

[0007] The predicted values ​​of the renewable micro-sources in the microgrid within the next 24h period and the output limits of the energy storage equipment and the diesel engine are used as the optimization space of the decision variables of the operating cost and environmental cost objective functions, and the particle swarm is initialized in the optimization space;

[0008] Calculating the fitness value of the initialized particle swarm according to the operating cost and environmental cost objective functions, obtaining a non-dominated solution, and establishing an external archive REP according to the non-dominated solution;

[0009] According to the roulette strategy, the best particle leader in gbest is searched from the external archive REP;

[0010] Update the inertia factor and mutation rate according to the current number of iterations, perform PSO particle swarm operation, obtain and update the position and speed of the next generation of particles, update pbest, and update the external archive REP, and delete the poor particles when out of bounds to keep the external archive REP from overflowing;

[0011] Determine whether the number of iterations t reaches the maximum number of iterations. If the number of iterations t reaches the maximum number of iterations t max , then output the external archive REP, which is the Pareto optimal solution set found;

[0012] A compromise solution with the minimum operating cost and environmental cost is found in the Pareto optimal solution set as the microgrid next-day scheduling strategy solution.

[0013] In one embodiment, the method further comprises: using the predicted values ​​of each micro-source in each time period of the next day in the microgrid and the output limits of the energy storage device and the diesel engine as the optimization space of the decision variables of the operating cost and environmental cost objective function, and before initializing the particle swarm in the optimization space, the method further comprises:

[0014] Obtain the first parameter information, and predict the photovoltaic power generation within the next 24h period according to the first parameter information and the LSTM model to obtain the photovoltaic prediction value P within the next 24h period PV (t), t∈{1,2,…24};

[0015] Obtain the second parameter information, and predict the wind turbine power generation within the next 24h period based on the second parameter information and the LSTM model to obtain the wind power prediction value P within the next 24h period wt (t), t∈{1,2,…24};

[0016] Get the diesel engine output limit P within the next 24 hours DE (t), t∈{1,2,…24} and the energy storage output limit P in the next 24h period BS (t), t∈{1,2,…24};

[0017] The optimization space of taking the predicted values ​​of each micro-source in each time period of the next day in the microgrid and the output limit of the energy storage device and the diesel engine as the decision variables of the operating cost and environmental cost objective function, and initializing the particle swarm in the optimization space includes:

[0018] Set the diesel engine output limit value P for the next 24h period DE(t), t∈{1,2,…24}, the energy storage output limit value P within the next 24h period BS (t), t∈{1,2,…24}, the photovoltaic prediction value P within the next 24h period PV (t), t∈{1,2,…24} and the wind power forecast value P for the next 24h period wt (t), t∈{1,2,…24} is used as the optimization space of decision variables of the operating cost and environmental cost objective functions, and 100 particle swarms with a scale of 96 decision variables are initialized.

[0019] In one embodiment, the first parameter information includes at least one of irradiance, temperature information, wind speed information and inverter power generation; the second parameter information includes at least one of wind speed and wind turbine power generation.

[0020] In one embodiment, the method comprises:

[0021] If the number of iterations t does not reach the maximum number of iterations t max , then let t=t+1, and return to search for the best particle leader in gbest from the external archive REP according to the roulette strategy.

[0022] In one embodiment, the method of searching for a compromise solution with the minimum operating cost and environmental cost in the Pareto optimal solution set as a microgrid next-day scheduling strategy includes:

[0023] The operating cost of each non-dominated solution in the external archive REP is normalized and added to the environmental cost function value to obtain the total cost of each non-dominated solution;

[0024] The non-dominated solution with the minimum total cost is determined as the optimal compromise solution of the microgrid operating cost and the environmental cost found by the multi-objective optimization algorithm, and the optimal compromise solution is used as the microgrid next-day scheduling strategy solution.

[0025] According to a second aspect of an embodiment of the present disclosure, a microgrid optimization scheduling device based on a multi-objective particle swarm optimization algorithm is provided, the device comprising: an initialization module, a calculation module, a search module, an update module, a judgment module and a query module; wherein,

[0026] The initialization module is used to use the predicted values ​​of the renewable micro-sources in the microgrid within the next 24h period and the output limits of the energy storage equipment and the diesel engine as the optimization space of the decision variables of the operating cost and environmental cost objective functions, and initialize the particle swarm in the optimization space;

[0027] The calculation module is used to calculate the fitness value of the initialized particle swarm according to the operating cost and environmental cost objective function, obtain a non-dominated solution, and establish an external archive REP according to the non-dominated solution;

[0028] The search module is used to search for the best particle leader in gbest from the external archive REP according to the roulette strategy;

[0029] The update module is used to update the inertia factor and mutation rate according to the current number of iterations, perform PSO particle swarm operations, obtain and update the position and speed of the next generation of particles, update pbest, and update the external archive REP, and delete the poor particles when out of bounds to keep the external archive REP from overflowing;

[0030] The judgment module is used to judge whether the number of iterations t reaches the maximum number of iterations. max , then output the external archive REP, which is the Pareto optimal solution set found;

[0031] The query module is used to use the compromise solution with the minimum operating cost and environmental cost queried from the Pareto optimal solution set as the microgrid next day scheduling strategy solution.

[0032] In one embodiment, the device further includes: a first acquisition module, a second acquisition module and a third acquisition module; wherein,

[0033] The first acquisition module is used to acquire the first parameter information, and predict the photovoltaic power generation within the 24h period of the next day according to the first parameter information and the LSTM model, to obtain the photovoltaic prediction value P within the 24h period of the next day PV (t), t∈{1,2,…24};

[0034] The second acquisition module is used to obtain the second parameter information, and predict the wind turbine power generation within the 24h period of the next day according to the second parameter information and the LSTM model to obtain the wind power prediction value P within the 24h period of the next day wt (t), t∈{1,2,…24};

[0035] The third acquisition module is used to obtain the diesel engine output limit value P within the next 24h period. DE (t), t∈{1,2,…24} and the energy storage output limit P in the next 24h period BS (t),t∈{1,2,…24}.

[0036] In one embodiment, the first parameter information includes at least one of irradiance, temperature information, wind speed information and inverter power generation; the second parameter information includes at least one of wind speed and wind turbine power generation.

[0037] In one embodiment, the device further comprises: a return module; wherein,

[0038] The return module is used to return to the state when the number of iterations t does not reach the maximum number of iterations t max When t=t+1, return to search for the best particle leader in gbest from the external archive REP according to the roulette strategy.

[0039] The query module includes: a normalization submodule and a determination submodule; wherein,

[0040] The normalization submodule is used to perform normalized addition processing on the operating cost of each non-dominated solution in the external archive REP and the environmental cost function value to obtain the total cost of each non-dominated solution;

[0041] The determination submodule is used to determine the non-dominated solution with the minimum total cost as the compromise optimal solution of the microgrid operating cost and the environmental cost found by the multi-objective optimization algorithm, and use the compromise optimal solution as the microgrid next day scheduling strategy solution.

[0042] The microgrid optimization scheduling method based on a multi-objective particle swarm optimization algorithm provided by the embodiment of the present disclosure has the following technical effects:

[0043] Accurate prediction: Through power prediction technology, it is possible to accurately predict the power generation and load power of non-controllable renewable energy in the short term in the future, providing accurate data support for optimized scheduling.

[0044] Multi-objective optimization: Comprehensively consider the two objectives of operating cost and environmental cost to achieve dual optimization of the economy and environmental protection of the microgrid.

[0045] Dynamic optimization: Using the MOPSO algorithm for dynamic optimization can quickly find the optimal scheduling solution that meets multiple objectives.

[0046] Strong adaptability: Optimization models are constructed for different types of microgrids (grid-connected and off-grid), which have strong adaptability and versatility.

[0047] Short-cycle scheduling: Use a shorter scheduling cycle (1 hour) to improve the accuracy of power prediction and optimize scheduling.

[0048] In summary, the microgrid optimization scheduling method based on the MOPSO multi-objective particle swarm algorithm proposed in the present invention can effectively solve the problems existing in the traditional microgrid optimization scheduling method, improve the operating efficiency and environmental protection performance of the microgrid, and has important practical application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0050] Figure 1 It is a schematic diagram of a combined cooling, heating and power (CCHP) system of a gas turbine in an embodiment of the present disclosure;

[0051] Figure 2 A flowchart of a microgrid optimization scheduling method based on a multi-objective particle swarm optimization algorithm provided in an embodiment of the present disclosure;

[0052] Figure 3 A logic diagram of a microgrid optimization scheduling method based on a multi-objective particle swarm optimization algorithm provided in an embodiment of the present disclosure;

[0053] Figure 4 An architecture diagram of a microgrid optimization scheduling device based on a multi-objective particle swarm optimization algorithm provided in an embodiment of the present disclosure;

[0054] Figure 5 An architectural diagram of a microgrid optimization scheduling device based on a multi-objective particle swarm optimization algorithm provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0055] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are merely examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims.

[0056] The present invention proposes a microgrid optimization scheduling method based on MOPSO (multi-objective particle swarm optimization) algorithm, which predicts the non-controllable renewable energy power generation and load power within the future scheduling optimization period (1 hour) through power prediction, constructs operation cost and environmental cost optimization models for grid-connected microgrids and off-grid microgrids respectively, and performs multi-objective dynamic optimization based on the constraints under the grid-connected operation mode and the off-grid operation mode, and outputs the scheduling schemes of each micro-source for the next scheduling period.

[0057] The core of the present disclosure is to use the MOPSO algorithm for multi-objective optimization. The MOPSO algorithm simulates the behavior of particle swarms in nature and continuously updates the position and velocity of particles to search for the optimal solution of multiple objective functions. During the optimization process, the algorithm can simultaneously consider the two objectives of operating cost and environmental cost and find a set of balanced solutions (i.e., the Pareto optimal solution set). In addition, the present disclosure adopts a shorter scheduling cycle (1 hour) to improve the accuracy of power prediction and optimization scheduling.

[0058] The models of each micro-source involved in the present disclosure will be described in detail below.

[0059] 1. Solar Photovoltaic Cell (PV) Mathematical Model:

[0060] 1.1 The common DC bus microgrid is as follows:

[0061] The output power of photovoltaic cells is related to the light intensity. Photovoltaic cells generally work in the maximum power point tracking (MPPT) mode, and their output power can be expressed as:

[0062] P pv =ζη m A p η p cosθ;

[0063] 1.2 The common AC bus microgrid is as follows:

[0064] The actual output power of the photovoltaic array can be obtained from the output power under standard rated conditions, light intensity, and ambient temperature. The output power can be expressed as:

[0065]

[0066] Where: P pv is the output power at the operating point;

[0067] P STC It is the operating parameters under standard test conditions (STC);

[0068] G STC is the solar irradiance, which is 1kW / m;

[0069] T STC The cell surface temperature is 25°C and the relative atmospheric optical quality is AM1.5.

[0070] G C is the irradiance at the working point, which is the radiation energy projected onto a unit area per unit time;

[0071] k is the power temperature coefficient, also known as the Boltzmann coefficient, and its value is 1.3806505×10 -23 J / K, which expresses the relationship between temperature and energy;

[0072] P STC is the rated output power of the photovoltaic array under standard rated conditions. If nPV is the total number of photovoltaic cells in the photovoltaic array, P STC is the rated output power of the photovoltaic cell, then P STC =nPV*P STC ; Inverse temperature characteristics - the possibility of inverter tripping at noon in winter is much higher than in the afternoon in summer, and the load on the inverter is higher.

[0073] T C is the battery surface temperature at the operating point, which is the ambient temperature T a And the function of wind speed:

[0074] T C =T a +αG C ;

[0075] Where: The coefficient α is an exponential function of the wind speed v:

[0076]

[0077] 2. Wind Turbine (WT) mathematical model:

[0078] The approximate relationship between the output power of a wind turbine and wind speed can be expressed by the following piecewise function:

[0079]

[0080] Where: v ci is the cut-in wind speed; v co is the cut-out wind speed; v r is the rated wind speed; P r is the rated output power of the wind turbine.

[0081] When the wind speed is between v ci and v r When the wind turbine output power is between , the wind speed function η(v) can be expressed as a linear relationship, that is:

[0082] η(v)=P r (v 3 -v c i3 )(v r3 -v ci3 );

[0083] When testing the algorithm, vci 、v co 、v r The speeds can be 3m / s, 25m / s and 14m / s respectively.

[0084] 3. Micro Turbine (MT) mathematical model:

[0085] Micro gas turbines use natural gas, biogas and other fuels to provide heat and electricity. The schematic diagram of the combined cooling, heating and power (CCHP) system of the gas turbine is as follows: Figure 1 As shown. The high temperature and more available energy are used to generate electricity, while the relatively low temperature is cooled or heated by the lithium bromide unit, which improves the energy utilization rate and helps to improve economic and environmental benefits.

[0086] If there is a controllable micro gas turbine in the microgrid, its fuel cost is related to its own efficiency. The following mathematical model is constructed using the C65 gas turbine as an example. The fuel cost expression of MT in time interval j is:

[0087]

[0088] In the formula, and It is a cubic function relationship:

[0089]

[0090] Where: C nl is the natural gas price, 2 yuan / m 3 ;

[0091] LHV f The low calorific value of natural gas is 9.7 kWh / m 3 ;

[0092] is the net output power in time interval j;

[0093] is the net power generation efficiency of the gas turbine in time interval j.

[0094] 4. Diesel Generator (DE) mathematical model:

[0095] The fuel cost of DE is its consumption characteristic function. The fuel cost of DE is expressed as a quadratic function:

[0096] C DE =α+βP DE +γP DE 2 ;

[0097] Where: CDE is the fuel cost of DE;

[0098] P DE is the output power of DE;

[0099] α, β, and γ are the fuel cost coefficients of DE, α=6, β=0.12, and γ=8.5×10 -4 .

[0100] 5. Mathematical model of energy storage system (ESS):

[0101] The battery can track the changes in wind and solar power output to charge and discharge, which plays a role in buffering the uncertain output of wind and solar power in the power grid, improving the reliability and continuity of the power supply of the power grid. When the total output power of DG is greater than the total load, ES charges, otherwise, ES discharges. The charge and discharge state of the battery is expressed as:

[0102]

[0103] Among them, E SB (t), E SB (t-1) are the capacities of the battery in period t and period t-1 respectively;

[0104] P total (t) is the sum of the outputs of all micro-sources during period t;

[0105] P load (t) is the total load of the system during period t; η inv The working efficiency of the converter;

[0106] η sb is the charge and discharge efficiency of ES.

[0107] The multi-objective mathematical model for optimal operation of microgrids will be described in detail below.

[0108] Off-grid microgrid: The microgrid is in an island operation mode. Under the conditions of meeting system constraints, the economy, reliability and environmental protection of the microgrid are comprehensively considered to establish a multi-objective optimization scheduling model for the microgrid with the lowest operating cost and pollutant emission cost of the microgrid system.

[0109] Operating cost: The power generation cost of a microgrid mainly considers the operating cost and the compensation cost of the removable load. Therefore, the power generation cost satisfies the system equality constraints and inequality constraints, so that the total operating cost of N dispatchable micro-power sources and energy storage devices in the microgrid in T time periods is minimized. The dispatch period T usually takes values ​​of 24 hours, 1 month, 1 quarter, 1 year, etc. The principle of selecting the total time period within the dispatch period is to ensure that the power distribution in the microgrid remains basically unchanged within a time period, that is, to meet the premise that the output power of the microgrid and the load demand power do not change much. In this way, the microgrid can be considered to be in a stable state during analysis and calculation, without considering its transient process. That is:

[0110]

[0111] Where: T is the number of time periods in the dispatch cycle of the microgrid;

[0112] N is the number of micro power types;

[0113] CO i P i (t) is the total power generation cost of micro-source i in period t;

[0114] IR t is the interruptible cost of the microgrid in period t;

[0115] IR(t) is the unit power generation cost of micro-source i in period t;

[0116] P i (t) is the power generation of micro power source i in period t;

[0117] The operating cost in the above formula is equal to the sum of the total power generation cost and the cost of removable loads.

[0118] The total electricity generation cost is calculated as follows:

[0119] CO i P i (t) = C i P i (t)+IV i P i (t)+OM i P i (t);

[0120] Where: C i The fuel cost consumed per unit power generated by micro power source i, unit: yuan / kW;

[0121] IV i is the depreciation cost of micro power source i converted to unit power generation, unit: yuan / kW;

[0122] OMi is the maintenance cost per unit power of micro power source i, unit: yuan / kW;

[0123] C i P i (t) is the fuel cost of micro power source i in period t;

[0124] IV i P i (t) is the depreciation cost of micro-source i in period t;

[0125] OM i P i (t) is the maintenance cost of micro power source i in period t.

[0126] The calculation formula for depreciation cost per unit of power generation is as follows:

[0127]

[0128] Where: C INS,i is the installation cost of micro power source i;

[0129] P r,i is the rated power of micro power source i;

[0130] f c,i is the capacity factor of micro power source i;

[0131] d i is the depreciation rate of micro power source i;

[0132] m i The service life of the micro power supply i.

[0133] The calculation formula of maintenance cost per unit power generation is as follows:

[0134] OM i,t =K m,i ;

[0135] Where: K m,i It is the maintenance cost of unit power generation of micro power source i.

[0136] Load shedding costs: When the microgrid is operating in an isolated island, there is a situation of insufficient power supply, and some non-important loads need to be shelved to ensure the normal power supply of important loads. The strategy for shelving some loads is: only when the controllable power supply, uncontrollable power supply and energy storage device are all in full power state, the load with insufficient power supply will be shelved. At the same time, corresponding economic compensation needs to be made for the shelved secondary loads, that is:

[0137] IR(t)=A+BP IL +CP IL 2 ;

[0138] Where: A, B, and C are the load removal coefficients of the microgrid, which can be A=6.14, B=1.2, and C=1.23×10 -4 , P IL is the power removed from the microgrid.

[0139] Environmental cost: Environmental cost mainly considers CO2, SO2, and NO x The emission treatment cost. Since PV and WT are clean energy and do not produce polluting gases during operation, the environmental cost of PV and WT is not considered. The objective function is to minimize the environmental cost, and its expression is:

[0140]

[0141] Among them: F env for environmental costs;

[0142] K is the type of pollutant emitted (CO2, SO2 and NO x );

[0143] α j is the unit cost of treating the jth pollutant, unit: yuan / kg;

[0144] β i,j It is the emission coefficient of the jth pollutant emitted when the micro power source i outputs unit power, unit: g / kW;

[0145] β i,j P(t) i The power generated by micro power source i in time period t is P(t) i When , the mass of the jth type of pollution generated;

[0146] is the mass of the jth type of pollution generated by all micro-power sources in time period t;

[0147] The cost of treating the jth type of pollution for all micro-power sources in time period t;

[0148] The cost of handling all pollution for all micro-sources in time period t;

[0149] The cost of all pollution for processing all micro-power sources in time period T.

[0150] The operating parameters of each micro-source in the microgrid are shown in Table 1.

[0151] Table 1: Operating parameters of each micro-source in the microgrid

[0152]

[0153] Pollutant emissions and treatment parameters are shown in Table 2 below.

[0154] Table 2: Pollutant emissions and treatment parameters

[0155]

[0156]

[0157] The energy storage parameters are shown in Table 3 below.

[0158] Table 3: Energy storage parameters

[0159]

[0160] The constraints are as follows:

[0161] The output constraints of micro power sources are as follows:

[0162] P i,min ≤P i ≤P i,max ;

[0163] Where: P i,min , P i,max They are respectively the upper and lower limits of the micro power source output.

[0164] The power balance constraints of the microgrid are as follows:

[0165] P PV (t)+P WT (t)+P DE (t)+P MT (t)+P BS (t) = P L (t)+P IL (t);

[0166] Where: P PV (t) is the photovoltaic power generation in time period t;

[0167] P WT (t) is the power generated by the wind turbine in time period t;

[0168] P DE (t) is the power generated by the diesel generator in time period t;

[0169] P MT (t) is the power generated by the gas turbine in time period t;

[0170] P BS (t) is the charging and discharging power of the energy storage in the time period t. BS (t)>0 indicates discharge, when P BS(t)<0 means charging;

[0171] P L (t) is the total load of the microgrid in time period t;

[0172] P IL (t) is the load removed by the microgrid during time period t.

[0173] The microgrid line transmission power constraints are as follows:

[0174] P line,min ≤P line ≤P line,max ;

[0175] Where: P line,min , P line,max They are the upper and lower limits of the line transmission power of the microgrid in time period t.

[0176] The battery operation constraints are as follows:

[0177] P BS,min ≤P BS ≤P BS,max ;

[0178] E BS,min ≤E BS ≤E BS,max ;

[0179] Where: P BS,min , P BS,max are the minimum and maximum charging and discharging power of the energy storage system respectively;

[0180] E BS,min 、E BS,max are the minimum and maximum capacities of the energy storage system respectively;

[0181] The MT output ramp constraints are as follows:

[0182] P MT (t)-P MT (t-1)|≤R;

[0183] Where: R is the limit for increasing or decreasing the power of the micro gas turbine.

[0184] The microgrid optimization scheduling method based on the multi-objective particle swarm optimization algorithm provided by the embodiment of the present disclosure will be described in detail below.

[0185] Figure 2 Flow chart of a microgrid optimization scheduling method based on a multi-objective particle swarm optimization algorithm provided in an embodiment of the present disclosure. Figure 2 As shown, the method includes:

[0186] Step 201: Taking the predicted values ​​of each renewable micro-source (photovoltaic, wind power, etc.) in the microgrid within the next 24h period and the output limits of the energy storage equipment and the diesel engine as the optimization space of the decision variables of the operating cost and environmental cost objective functions, and initializing the particle swarm in the optimization space;

[0187] In one embodiment, the predicted values ​​of each renewable micro-source (photovoltaic, wind power, etc.) in the microgrid within the next 24h period and the output limits of the energy storage device and the diesel engine are used as the optimization space of the decision variables of the operating cost and environmental cost objective function, and before the particle swarm is initialized in the optimization space, the method further includes:

[0188] Obtain the first parameter information, and predict the photovoltaic power generation within the next 24h period according to the first parameter information and the LSTM model to obtain the photovoltaic prediction value P within the next 24h period PV (t), t∈{1,2,…24};

[0189] Obtain the second parameter information, and predict the wind turbine power generation within the next 24h period based on the second parameter information and the LSTM model to obtain the wind power prediction value P within the next 24h period wt (t), t∈{1,2,…24};

[0190] Get the diesel engine output limit P within the next 24 hours DE (t), t∈{1,2,…24} and the energy storage output limit P in the next 24h period BS (t), t∈{1,2,…24};

[0191] The optimization space of taking the predicted values ​​of each renewable micro-source (photovoltaic, wind power, etc.) in the microgrid within the next 24h period and the output limit of the energy storage device and the diesel engine as the decision variables of the operating cost and environmental cost objective function, and initializing the particle swarm in the optimization space includes:

[0192] Set the diesel engine output limit value P for the next 24h period DE (t), t∈{1,2,…24}, the energy storage output limit value P within the next 24h period BS (t), t∈{1,2,…24}, the photovoltaic prediction value P within the next 24h period PV (t), t∈{1,2,…24} and the wind power forecast value P for the next 24h period wt (t), t∈{1,2,…24} is used as the optimization space of decision variables of the operating cost and environmental cost objective functions, and 100 particle swarms with a scale of 96 decision variables are initialized.

[0193] The first parameter information includes at least one of irradiance, temperature information, wind speed information and inverter power generation; the second parameter information includes at least one of wind speed and wind turbine power generation.

[0194] In this embodiment, based on the load history data, meteorological data, real-time electricity price, working days / weekends, etc., the LSTM model is used to predict the load for the next 24 hours, and the load prediction value P for each hour of the next 24 hours is obtained. L (t), t∈{1,2,…24}; Based on irradiance, temperature, wind speed, inverter power generation, etc., the LSTM model is used to predict the photovoltaic power generation for the next 24 hours, and the photovoltaic prediction value P for each hour of the next 24 hours is obtained. PV (t), t∈{1,2,…24}; Based on wind speed and wind turbine power generation, the LSTM model is used to predict the wind turbine power generation for the next 24 hours, and the hourly wind power forecast value P for the next 24 hours is obtained. wt (t),t∈{1,2,…24}.

[0195] It should be noted that the micro-sources in this embodiment include photovoltaic, wind turbine, diesel engine, and energy storage. The predicted value of each micro-source in the microgrid for each period of 24 hours the next day is used as the 96 decision variables of the objective function. The photovoltaic predicted value P PV (t), t∈{1,2,…24}, wind power prediction value P wt (t), t∈{1,2,…24}, diesel engine output limit P DE (t), t∈{1,2,…24}, and the energy storage output limit P BS (t), t∈{1,2,…24} is used as the decision variable optimization space for the operating cost and environmental cost objective functions. 100 particle swarms (position and speed) and parameters with a scale of 96 decision variables are initialized, and the boundary conditions of the 96 decision variables are initialized according to the system constraints and predicted values ​​in this disclosure. The upper and lower limits of the particles are set for photovoltaic and wind power according to the predicted values, the upper and lower limits are set for diesel engines according to the minimum and maximum output powers, the upper and lower limits are set for the maximum input and output powers of energy storage, and the initialization algebra t=1 is set.

[0196] Step 202: Calculate the fitness value of the initialized particle swarm according to the operating cost and environmental cost objective functions, obtain a non-dominated solution, and establish an external archive REP according to the non-dominated solution;

[0197] The fitness values ​​of the 100 initialized particles are calculated according to the operating cost and environmental cost objective functions, and the non-dominated solutions are sorted to obtain the non-dominated solutions, and an external archive REP is established, which contains all candidate non-dominated solutions:

[0198] REP=Repositorye(swarm,rep_size,grid_size,alpha,beta,gama);

[0199] Step 203: searching for the best particle leader in gbest from the external archive REP according to the roulette strategy;

[0200] In this step, the gbest best particle leader is selected from REP according to the roulette strategy:

[0201] leader=REP.Selectleader();

[0202] Step 204: update the inertia factor and mutation rate according to the current number of iterations, perform PSO particle swarm operation, obtain and update the position and velocity of the next generation of particles, update pbest, and update the external archive REP, and delete the poor particles when out of bounds to keep the external archive REP from overflowing;

[0203] In this step, the inertia factor and mutation rate are updated according to the current number of iterations, the PSO particle swarm operation is performed, the position and velocity of the next generation of particles are updated, the particle swarm Pt+1 generation performs a dominance relationship comparison, and updates pbest:

[0204] swarm(i)=swarm(i).update(wc,c,pc,leader,problem);

[0205] Update the external repository REP, delete the poor particles when out of bounds, and keep REP from overflowing;

[0206] ERP=ERP.update(swarm)

[0207] Step 205: Determine whether the number of iterations t reaches the maximum number of iterations. If the number of iterations t reaches the maximum number of iterations t max , then output the external archive REP, which is the Pareto optimal solution set found;

[0208] In this step, it is determined whether the maximum number of iterations has been reached. If so, the external archive REP is output, and the external archive REP is the Pareto optimal solution set found.

[0209] In one embodiment, the method comprises:

[0210] If the number of iterations t does not reach the maximum number of iterations t max, then let t=t+1, and return to search for the best particle leader in gbest from the external archive REP according to the roulette strategy.

[0211] In this step, if the maximum number of iterations t is not reached max , then let t=t+1, and return to search for the best particle leader in gbest from the external archive REP according to the roulette strategy.

[0212] Step 206: Find a compromise solution with the minimum operating cost and environmental cost in the Pareto optimal solution set as the microgrid next day scheduling strategy solution.

[0213] In one embodiment, the method of searching for a compromise solution with the minimum operating cost and environmental cost in the Pareto optimal solution set as a microgrid next-day scheduling strategy includes:

[0214] The operating cost of each non-dominated solution in the external archive REP is normalized and added to the environmental cost function value to obtain the total cost of each non-dominated solution;

[0215] The non-dominated solution with the minimum total cost is determined as the optimal compromise solution of the microgrid operating cost and the environmental cost found by the multi-objective optimization algorithm, and the optimal compromise solution is used as the microgrid next-day scheduling strategy solution.

[0216] In this embodiment, the operating cost and environmental cost function value of each non-dominated solution in the external storage repository REP are normalized and added to obtain the total cost of each non-dominated solution. The group of non-dominated solutions with the smallest total cost is the optimal solution with the smallest microgrid operating cost and environmental cost found by the MOPSO multi-objective optimization algorithm, that is, the wind, solar, diesel and micro-source dispatching output plan in the microgrid for 24 hours the next day.

[0217] Figure 3 A logic diagram of a microgrid optimization scheduling method based on a multi-objective particle swarm optimization algorithm provided in an embodiment of the present disclosure.

[0218] The disclosed embodiment provides a microgrid optimization scheduling method based on a multi-objective particle swarm optimization algorithm. Through power prediction technology, it can accurately predict the power generation and load power of non-controllable renewable energy in the short future period, and provide accurate data support for optimization scheduling. By comprehensively considering the two goals of operating cost and environmental cost, it can achieve dual optimization of the economy and environmental protection of the microgrid. By using the MOPSO algorithm for dynamic optimization, it can quickly find the optimal scheduling method that meets multiple goals. By constructing optimization models for different types of microgrids (grid-connected and off-grid), it has strong adaptability and versatility. By adopting a shorter scheduling cycle (1 hour), the accuracy of power prediction and the accuracy of optimization scheduling are improved.

[0219] In summary, the microgrid optimization scheduling method based on the MOPSO multi-objective particle swarm algorithm proposed in the present invention can effectively solve the problems existing in the traditional microgrid optimization scheduling method, improve the operating efficiency and environmental protection performance of the microgrid, and has important practical application value.

[0220] Figure 4 The following is an architecture diagram of a microgrid optimization scheduling device based on a multi-objective particle swarm optimization algorithm provided in an embodiment of the present disclosure. Figure 4 As shown, the device includes: an initialization module 401, a calculation module 402, a search module 403, an update module 404, a judgment module 405 and a query module 406; wherein the initialization module 401 is used to use the predicted values ​​of the renewable micro-sources in the microgrid within the 24h period of the next day and the output limits of the energy storage equipment and the diesel engine as the optimization space of the decision variables of the operating cost and environmental cost objective function, and initialize the particle swarm in the optimization space; the calculation module 402 is used to calculate the fitness value of the initialized particle swarm according to the operating cost and environmental cost objective function, obtain a non-dominated solution, and according to the non-dominated solution The dominant solution establishes an external archive REP; the search module 403 is used to search for the best particle leader in gbest from the external archive REP according to the roulette strategy; the update module 404 is used to update the inertia factor and mutation rate according to the current number of iterations, perform PSO particle swarm operations, obtain and update the position and speed of the next generation of particles, update pbest, and update the external archive REP, and delete the poor particles when out of bounds to keep the external archive REP from overflowing; the judgment module 405 is used to judge whether the iteration number t reaches the maximum number of iterations. If the iteration number t reaches the maximum number of iterations t max , then output the external archive REP, the external archive REP is the Pareto optimal solution set found; the query module 406 is used to use the compromise solution with the minimum operating cost and environmental cost found in the Pareto optimal solution set as the microgrid next day scheduling strategy solution.

[0221] Figure 5 The following is an architecture diagram of a microgrid optimization scheduling device based on a multi-objective particle swarm optimization algorithm provided in an embodiment of the present disclosure. Figure 5 As shown, the device includes: a first acquisition module 501, a second acquisition module 502, a third acquisition module 503, an initialization module 504, a calculation module 505, a search module 506, an update module 507, a judgment module 508 and a query module 509; wherein the first acquisition module 501 is used to obtain first parameter information, and predict the photovoltaic power generation within the 24h period of the next day according to the first parameter information and the LSTM model, and obtain the photovoltaic prediction value P within the 24h period of the next dayPV (t), t∈{1,2,…24}; The second acquisition module 502 is used to obtain the second parameter information, and predict the wind turbine power generation within the next 24h period according to the second parameter information and the LSTM model, and obtain the wind power prediction value P within the next 24h period wt (t), t∈{1,2,…24}; the third acquisition module 503 is used to obtain the diesel engine output limit value P within the 24h period of the next day; DE (t), t∈{1,2,…24} and the energy storage output limit P in the next 24h period BS (t),t∈{1,2,…24}.

[0222] In one embodiment, the first parameter information includes at least one of irradiance, temperature information, wind speed information and inverter power generation; the second parameter information includes at least one of wind speed and wind turbine power generation.

[0223] In one embodiment, the device further comprises: a return module; wherein,

[0224] The return module is used to return to the state when the number of iterations t does not reach the maximum number of iterations t max When t=t+1, return to search for the best particle leader in gbest from the external archive REP according to the roulette strategy.

[0225] In one embodiment, the query module includes: a normalization submodule and a determination submodule; wherein,

[0226] The normalization submodule is used to perform normalized addition processing on the operating cost of each non-dominated solution in the external archive REP and the environmental cost function value to obtain the total cost of each non-dominated solution;

[0227] The determination submodule is used to determine the non-dominated solution with the minimum total cost as the compromise optimal solution of the microgrid operating cost and the environmental cost found by the multi-objective optimization algorithm, and use the compromise optimal solution as the microgrid next day scheduling strategy solution.

[0228] Those skilled in the art can understand that all or part of the steps of implementing the above-mentioned method embodiments can be completed by hardware related to program instructions. The above-mentioned program can be pre-stored in a computer-readable storage medium. When the program is executed, the steps of the above-mentioned method embodiments are executed; and the above-mentioned storage medium includes: ROM, RAM, disk or optical disk and other media that can store program codes.

[0229] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the disclosure disclosed herein. The present disclosure is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The description and examples are to be considered exemplary only, and the true scope and spirit of the present disclosure are indicated by the following claims.

[0230] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. A microgrid optimization scheduling method based on a multi-objective particle swarm optimization algorithm, characterized in that: The method comprises: The predicted values ​​of each renewable micro-source in the microgrid within the next 24h period and the output limits of the energy storage equipment and the diesel engine are used as the optimization space of the decision variables of the operating cost and environmental cost objective functions, and the particle swarm is initialized in the optimization space; Calculating the fitness value of the initialized particle swarm according to the operating cost and environmental cost objective functions, obtaining a non-dominated solution, and establishing an external archive REP according to the non-dominated solution; According to the roulette strategy, the best particle leader in gbest is searched from the external archive REP; Update the inertia factor and mutation rate according to the current number of iterations, perform PSO particle swarm operation, obtain and update the position and speed of the next generation of particles, update pbest, and update the external archive REP, and delete the poor particles when out of bounds to keep the external archive REP from overflowing; Determine whether the number of iterations t reaches the maximum number of iterations. If the number of iterations t reaches the maximum number of iterations t max , then output the external archive REP, which is the Pareto optimal solution set found; A compromise solution with the minimum operating cost and environmental cost is found in the Pareto optimal solution set as the microgrid next-day scheduling strategy solution.

2. The method according to claim 1, characterized in that The method further comprises: using the predicted values ​​or output limits of each renewable micro-source and energy storage device in the microgrid and the diesel engine in each time period of the next day as the decision variable optimization space of the operation cost and environmental cost objective function, and before initializing the particle swarm in the optimization space, the method further comprises: Obtain the first parameter information, and predict the photovoltaic power generation within the next 24h period according to the first parameter information and the LSTM model to obtain the photovoltaic prediction value P within the next 24h period PV (t), t∈{1,2,…24}; Obtain the second parameter information, and predict the wind turbine power generation within the next 24h period based on the second parameter information and the LSTM model to obtain the wind power prediction value P within the next 24h period wt (t), t∈{1,2,…24}; Get the diesel engine output limit P within the next 24 hours DE (t), t∈{1,2,…24} and the energy storage output limit P in the next 24h period BS (t), t∈{1,2,…24}; The method of taking the predicted values ​​or output limits of each renewable micro-source and energy storage device in the microgrid and the diesel engine in each time period of the next day as the decision variable optimization space of the operating cost and environmental cost objective function, and initializing the particle swarm includes: Set the diesel engine output limit value P for the next 24h period DE (t), t∈{1,2,…24}, the energy storage output limit value P within the next 24h period BS (t), t∈{1,2,…24}, the photovoltaic prediction value P within the next 24h period PV (t), t∈{1,2,…24} and the wind power forecast value P for the next 24h period wt (t), t∈{1,2,…24} is used as the optimization space of decision variables of the operating cost and environmental cost objective functions, and 100 particle swarms with a scale of 96 decision variables are initialized.

3. The method according to claim 2, characterized in that The first parameter information includes at least one of irradiance, temperature information, wind speed information and inverter power generation; the second parameter information includes at least one of wind speed and wind turbine power generation.

4. The method according to any one of claims 1 to 3, characterized in that: The method comprises: If the number of iterations t does not reach the maximum number of iterations t max , then let t=t+1, and return to search for the best particle leader in gbest from the external archive REP according to the roulette strategy.

5. The method according to any one of claims 1 to 4, characterized in that: The compromise solution with the minimum operating cost and environmental cost is queried in the Pareto optimal solution set as the microgrid next day dispatch strategy solution, which includes: The operating cost of each non-dominated solution in the external archive REP is normalized and added to the environmental cost function value to obtain the total cost of each non-dominated solution; The non-dominated solution with the minimum total cost is determined as the optimal compromise solution of the microgrid operating cost and the environmental cost found by the multi-objective optimization algorithm, and the optimal compromise solution is used as the microgrid next-day scheduling strategy solution.

6. A microgrid optimization dispatching device based on a multi-objective particle swarm optimization algorithm, characterized in that: The device comprises: an initialization module, a calculation module, a search module, an update module, a judgment module and a query module; wherein, The initialization module is used to use the predicted values ​​of each renewable micro-source in the microgrid within the next 24h period and the output limits of the energy storage device and the diesel engine as the optimization space of the decision variables of the operating cost and environmental cost objective function, and initialize the particle swarm in the optimization space; The calculation module is used to calculate the fitness value of the initialized particle swarm according to the operating cost and environmental cost objective function, obtain a non-dominated solution, and establish an external archive REP according to the non-dominated solution; The search module is used to search for the best particle leader in gbest from the external archive REP according to the roulette strategy; The update module is used to update the inertia factor and mutation rate according to the current number of iterations, perform PSO particle swarm operations, obtain and update the position and speed of the next generation of particles, update pbest, and update the external archive REP, and delete the poor particles when out of bounds to keep the external archive REP from overflowing; The judgment module is used to judge whether the number of iterations t reaches the maximum number of iterations. max , then output the external archive REP, which is the Pareto optimal solution set found; The query module is used to use the compromise solution with the minimum operating cost and environmental cost queried from the Pareto optimal solution set as the microgrid next day scheduling strategy solution.

7. The device according to claim 6, characterized in that The device further includes: a first acquisition module, a second acquisition module and a third acquisition module; wherein, The first acquisition module is used to acquire the first parameter information, and predict the photovoltaic power generation within the 24h period of the next day according to the first parameter information and the LSTM model, to obtain the photovoltaic prediction value P within the 24h period of the next day PV (t), t∈{1,2,…24}; The second acquisition module is used to obtain the second parameter information, and predict the wind turbine power generation within the 24h period of the next day according to the second parameter information and the LSTM model to obtain the wind power prediction value P within the 24h period of the next day wt (t), t∈{1,2,…24}; The third acquisition module is used to obtain the diesel engine output limit value P within the next 24h period. DE (t), t∈{1,2,…24} and the energy storage output limit P in the next 24h period BS (t),t∈{1,2,…24}.

8. The device according to claim 7, characterized in that The first parameter information includes at least one of irradiance, temperature information, wind speed information and inverter power generation; the second parameter information includes at least one of wind speed and wind turbine power generation.

9. The device according to any one of claims 6 to 8, characterized in that The device further comprises: a return module; wherein, The return module is used to return to the state when the number of iterations t does not reach the maximum number of iterations t max When t=t+1, return to search for the best particle leader in gbest from the external archive REP according to the roulette strategy.

10. The device according to any one of claims 6 to 8, characterized in that The query module includes: a normalization submodule and a determination submodule; wherein, The normalization submodule is used to perform normalized addition processing on the operating cost of each non-dominated solution in the external archive REP and the environmental cost function value to obtain the total cost of each non-dominated solution; The determination submodule is used to determine the non-dominated solution with the minimum total cost as the compromise optimal solution of the microgrid operating cost and the environmental cost found by the multi-objective optimization algorithm, and use the compromise optimal solution as the microgrid next day scheduling strategy solution.

Images