Micro-grid cluster multi-target scheduling method based on self-adaptive multi-hybrid strategy bald eagle algorithm

Abstract

The invention discloses a micro-grid cluster multi-target scheduling method based on a self-adaptive multi-hybrid strategy bald eagle algorithm, which utilizes multi-micro-grid interconnection to realize electric energy coordination interaction and optimize internal resource distribution. On the premise that stable operation and installed capacity of multiple distributed power supplies are constrained, a micro-grid economic and environmental cost cluster multi-target scheduling model is constructed. And the management center determines optimization target parameters according to the obtained photovoltaic prediction data, basic load prediction data and real-time electricity price data of each micro-grid. And solving the scheduling model by using an improved Balde eagle algorithm which contains an adaptive adjustment factor and is combined with a multi-hybrid strategy to obtain an optimal solution and a solution set under the condition that sub-targets are minimum, reconstructing and optimizing a multi-objective function by using fuzzy membership satisfaction according to a result, and obtaining a power distribution scheme of the micro-grid cluster system. The method provided by the invention is higher in convergence speed and highest in precision, gives consideration to optimization between economic and environmental multiple targets, can effectively coordinate electric energy interaction between microgrids, and enables system operation to be more economical and environment-friendly.

Description

Technical Field

[0001] The present invention relates to the technical field of microgrid optimization scheduling, and in particular to a microgrid cluster multi-objective scheduling method based on an adaptive multi-hybrid strategy vulture algorithm. Background Art

[0002] In recent years, the sustainable development route of "green environmental protection, energy conservation and carbon reduction" has become a general consensus of the international community. Microgrids, as emerging micro-power systems, are used to supply power in remote areas. They can integrate multiple renewable energy sources while fully coordinating internal distributed power sources to ensure the green economic operation of the system. However, a single microgrid has great limitations in utilizing renewable energy and ensuring the safe and stable operation of micro-sources. A cluster system composed of multiple microgrids connected by interconnection lines can overcome the lack of flexibility, achieve energy coordination and interaction, optimize resource allocation between micro-power sources, meet the adaptability of microgrids to random changes in load characteristics, and improve the utilization efficiency of renewable energy. In view of the fact that the current optimization scheduling method cannot coordinate the power interaction between microgrids and balance the economic and environmental benefits of the operation of the microgrid cluster system, it is necessary to propose an optimization scheduling method applied to multi-microgrids to effectively solve the current problems.

[0003] At present, in the problem of optimal scheduling of multi-microgrid systems, the two main solution methods are operations research and optimization algorithm related methods. Operations research is mainly used to solve linear problems, integer linear programming and other problems. Faced with large-scale systems, the optimization objectives have nonlinear and multivariable characteristics, so this solution method is no longer applicable. The optimization algorithm includes an improved hybrid algorithm to solve the optimization problem of the microgrid system. Although it has stronger global search ability and faster solution speed than the traditional algorithm, its convergence speed is slower as the optimization object changes, and the result is easy to fall into the local optimum. Many non-multi-objective solution algorithms for multi-objective optimization problems fail to provide a feasible solution strategy. Therefore, it is very important to propose a microgrid cluster system optimization scheduling method that can effectively coordinate multiple optimization objects. Summary of the invention

[0004] In order to solve the above problems, the present invention provides a multi-objective scheduling method for microgrid clusters based on an adaptive multi-hybrid strategy vulture algorithm. A variety of hybrid strategies combined with adaptive variables give full play to the advantages of multiple parties and are applicable to complex system optimization problems, which can improve the convergence speed and accuracy. At the same time, the cluster system composed of multiple single microgrid systems can achieve coordinated interaction between microgrids, and can achieve flexible, economical and environmentally friendly operation of the cluster system through the principle of "more output and less input" and combined with the actual equipment operation status and goals. The fuzzy membership function performs normalized satisfaction evaluation on the sub-goals, thereby realizing multi-objective optimization scheduling.

[0005] The present invention is achieved through the following technical solutions:

[0006] A multi-objective scheduling method for microgrid clusters based on an adaptive multi-hybrid strategy vulture algorithm specifically includes the following steps:

[0007] Step S1: obtaining the day-ahead forecast output data of the photovoltaic units of the multi-microgrid system, and obtaining the day-ahead forecast data of the load within the system according to the internal load characteristics of the microgrid;

[0008] Step S2: Under the premise of stable operation of distributed power sources and installed capacity as constraints, a multi-objective scheduling model of economic and environmental costs is constructed by considering the power interaction between multiple microgrids;

[0009] Step S3: improving the traditional vulture algorithm based on adaptive variables combined with multi-optimization strategies, and using the improved algorithm to solve the microgrid cluster system scheduling model to obtain the optimal coordinated power and micro-source output solution;

[0010] Step S4: Based on the established environmental and economic multi-objective model, the economic and environmental objectives are remodeled using fuzzy membership satisfaction to construct a new fitness function, and a scheduling plan that can take into account both economic and environmental optimization is obtained. A comparative analysis is then conducted between the proposed scheduling strategy and the traditional scheduling strategy to evaluate the effectiveness of the optimization method.

[0011] Furthermore, in step S1, the day-ahead forecast data is obtained on a time scale of 1 hour, and the photovoltaic output and load of each microgrid in 24 hours have 24 data points respectively, for a total of 144 data points.

[0012] Furthermore, in step S2, power interaction refers to a method of exchanging power with other microgrids according to the principle of "more out and less in" based on the relationship between its own power generation and supply and load demand.

[0013] The objective function in the multi-objective optimization model is: Considering the microgrid cluster system under the interaction of multiple microgrids, the economic cost C during its operation should be comprehensively considered. ec and environmental cost C en . C ec Including power interaction Photovoltaic subsidies Micro power generation And equipment maintenance The economic operation objective function is:

[0014]

[0015] 1) Electricity interaction cost

[0016]

[0017] In the formula, σ buy and σ sell is the unit price of electricity purchased from the grid. It is the unit price of electricity purchased and sold between microgrids. is the interaction power between microgrid and large grid, represents the power exchange between microgrid n and microgrid m, and a and b are coefficients with values ​​of 0 and 1.

[0018] 2) Photovoltaic subsidy costs

[0019]

[0020] In the formula, Represents the unit price of photovoltaic subsidies, Represents the output of microgrid m at scheduling time t.

[0021] 3) Micro-power generation cost

[0022]

[0023] In the formula, is the photovoltaic replacement cost, is the total photovoltaic discharge capacity of microgrid m. ES and μ DE Respectively represent the operating status of the energy storage battery and DE, and σ DE They are ES conversion cost, diesel unit price and DE conversion cost. The output power of DE is The amount of fuel at Indicates the rated output power of the microgrid DE, f1 m is the fuel fitting factor.

[0024] 4) Equipment maintenance costs

[0025]

[0026] In the formula, K ES.om , K DE.om and K PV.om Represent the maintenance costs of energy storage batteries, diesel generators and photovoltaics per unit power respectively.

[0027] Objective function 2: Comprehensively consider the overall gas emissions to minimize environmental pollution

[0028] In the formula, C k represents the cost of treating different types of pollutants, γ DE.k and γ GRID.k They are the emission coefficients of different types of pollutants corresponding to diesel generators and distribution networks.

[0029] In step S3, in order to meet the needs of solving the optimization scheduling model, the BES algorithm performs the following improvement process:

[0030] S31. Population initialization;

[0031] S32. Improvement of spatial selection stage;

[0032] S33. Dive phase optimization;

[0033] S34.T distribution dynamic perturbations.

[0034] Further, the S31 is specifically: using Constructing the inverse solution The best individual in the combined population is used as the new generation of elite individuals Determine the new individual population through the elite reverse solution generation formula:

[0035]

[0036] Where, K∈rand(), α j and β j It is a dynamic boundary;

[0037] Further, the S32 is specifically as follows: select the LV strategy combined with the adaptive dynamic factor, define the step length formula as follows, and Alpha is dynamically adjusted by the algorithm iteration situation,

[0038]

[0039] In the early stage of optimization, a larger random step size is selected so that the algorithm can get rid of the interval where the local optimal solution is located. In the later stage, a smaller random step size is selected to search in the area near the optimal solution. The adaptation coefficient is as follows:

[0040] Alpha = w min +(w max -w min ) / 2*(1+cos(π*(iter-1) / (ger-1)))

[0041] In the formula, w min Take 5, w max Take 10, iter is the current iteration number, ger is the algorithm iteration number,

[0042] The control variable parameter α and the random parameter r are prone to fall into local optimality, so the LV flight strategy is introduced for updating. The corresponding spatial position update formula is:

[0043]

[0044] Set the threshold trail. If the global optimal fitness after the algorithm population is updated is worse than the global optimal fitness of the previous generation, the threshold trail+1. If the global optimal solution is better than the optimal solution of the previous iteration, the threshold trail=0. If the limit is exceeded when updating the trail during the optimization process, the LV update formula is used for space selection, otherwise the BES update formula is used for selection;

[0045] Furthermore, S33 is specifically as follows: using the strong global search capability of the golden sine strategy to achieve a balance between local and global searches, introducing the golden section coefficient to gradually narrow the search space and strengthen the search of the local space. The position transformation formula is:

[0046]

[0047] In the formula, represents the spatial position of the i-th individual at the t-th iteration, P i t =(P i,1 ,P i,2 ,...,P i,d ) represents the optimal position of the individual after iteration t. R1∈[0,2π] is a random number that determines the individual moving distance in the next iteration, R2∈[0,π] is a random number that determines the direction of the next movement, x1=-π+(1-τ)*2π, x2=-π+τ*2π are based on the golden ratio coefficients The irrational number obtained is

[0048] Further, the S34 is specifically as follows: adding T distribution disturbance to the global optimal individual, and adaptively updating the variable factor in combination with the iteration situation, to ensure the global search capability in the early stage of iteration and to give full play to the potential optimal solution mining capability of the algorithm in the later stage, in the form of P new,best =P best *(1+trnd(iter)).

[0049] Furthermore, in order to ensure the beneficial effect of the LV flight strategy selection, the maximum limit value in the trail update process is dynamically updated, and the dynamic variable adaptive change formula is as follows:

[0050] Wherein, the proportional coefficients of n1 and n2 to the upper and lower limits are set according to the optimization scenario. Generally, n1 is an integer between 5 and 10, and n2 is an integer between 40 and 80.

[0051] Furthermore, in step S4, considering the satisfaction of each sub-goal optimization goal, combined with the multi-objective model of step S2, the optimization objective function is reconstructed as follows:

[0052] Sat=min(min(exp((Cec_min -C ec ) / C ec_min ),exp((C en_min -C en ) / C en_min )),1).

[0053] Furthermore, the actual research results can be verified through actual microgrid operation cases, which are applicable to the economic and environmental multi-objective operation of all operating multi-microgrid systems within the included range.

[0054] The principle of this application is to apply the proposed target optimization strategy to the microgrid cluster system, and the optimization algorithm is improved on the basis of the traditional condor (BES) algorithm to form a solution method for the optimization scheduling model. In the multi-microgrid optimization, a mathematical model is constructed under full consideration of multi-microgrid coordination, and an economic and environmental multi-objective model is established. In the optimization of multiple objective functions, the optimal sub-objective is solved and a single objective equation combined with fuzzy satisfaction is constructed for system scheduling, and the multi-objective scheduling result is used as the power allocation plan for the microgrid cluster system and the interconnection line.

[0055] The beneficial effects of the present invention are:

[0056] 1. The improved vulture optimization algorithm based on adaptive variables combined with multi-hybrid strategies improves the traditional BES algorithm. It can significantly improve the convergence speed and convergence accuracy in solving the multi-objective optimization problem of cluster microgrids and find the global optimal solution.

[0057] 2. The fuzzy membership function evaluates the satisfaction of the sub-objective function and constructs a scheduling function that can take into account the optimal benefits of multiple economic and environmental objectives and achieve balanced optimization among multiple objectives.

[0058] 3. Through microgrid verification in actual application scenarios, this method can efficiently coordinate the power interaction between multiple microgrids, which is beneficial to the operation of the power grid and has great application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 This is the structure diagram of the microgrid cluster system.

[0060] Figure 2 Schematic diagram of the optimization process of the condor algorithm based on adaptive multi-hybrid strategies.

[0061] Figure 3 This is a data chart for photovoltaic output and load demand forecast.

[0062] Figure 4 The following is a convergence curve diagram comparing the optimization cost effects of different algorithms in an instance.

[0063] Figure 5 Interaction power diagram between microgrids after multi-objective optimization of fuzzy membership functions.

[0064] Figure 6 The output comparison diagrams of various micro-sources in different scenarios (6a, 6b, and 6c represent the comparison curves of microgrid MG1, MG2, and MG3, respectively). DETAILED DESCRIPTION

[0065] In order to make the method, purpose and technical solution of the present invention more clear, the present invention is further described in detail below with reference to the accompanying drawings and in combination with examples.

[0066] Example:

[0067] In this embodiment, Figure 1 As shown, the microgrid cluster system is composed of three microgrids, which are mainly composed of distributed micro-sources such as photovoltaic, diesel generator, energy storage battery and load. Each microgrid is connected by a tie line and there is power exchange between the microgrids and the large power grid. The energy management unit can realize the coordinated control of the power between microgrids, between the grid and the internal micro power supply unit, so that the multi-microgrid system can operate economically and environmentally friendly.

[0068] Consider using the improved vulture algorithm based on adaptive multi-hybrid strategy to perform multi-objective coordination optimization on microgrid clusters, including the following steps:

[0069] Step S1, obtaining the day-ahead predicted output data of the photovoltaic units of the multi-microgrid system, and obtaining the day-ahead predicted data of the load in the system according to the internal load characteristics of the microgrid;

[0070] In this embodiment, the day-ahead forecast data is obtained on a time scale of 1 hour, and the photovoltaic output and load of each microgrid system in 24 hours have 24 data points respectively, for a total of 144 data points.

[0071] Step S2, under the premise of stable operation of distributed power sources and installed capacity as constraints, a multi-objective scheduling model of economic and environmental costs is constructed by considering the power interaction between multiple microgrids;

[0072] The comprehensive cost of multi-microgrid system operation needs to consider the economic and environmental costs of its operation. The economic cost includes the energy interaction Photovoltaic subsidies Micro power generation And equipment maintenance The cost brought about. Economic cost objective function 1:

[0073]

[0074] (1) Electricity interaction cost

[0075]

[0076] In the formula, σ buy and σsell is the unit price of electricity purchased from the grid. It is the unit price of electricity purchased and sold between microgrids. is the interaction power between microgrid and large grid, represents the power exchange between microgrid n and microgrid m, and a and b are coefficients with values ​​of 0 and 1.

[0077] (2) PV subsidy costs

[0078]

[0079] In the formula, Represents the unit price of photovoltaic subsidies, Represents the output of microgrid m at scheduling time t.

[0080] (3) Micro-power generation cost

[0081]

[0082] In the formula, is the photovoltaic replacement cost, is the total photovoltaic discharge capacity of microgrid m. ES and μ DE Respectively represent the operating status of the energy storage battery and DE, and σ DE They are ES conversion cost, diesel unit price and DE conversion cost. The output power of DE is The amount of fuel at Indicates the rated output power of the microgrid DE, f1 m is the fuel fitting factor.

[0083] (4) Equipment maintenance costs

[0084] Environmental cost objective function 2: Comprehensively consider comprehensive gas emissions to minimize environmental pollution

[0085]

[0086] In the formula, C k represents the cost of treating different types of pollutants, γ DE.k and γ GRID.k They are the emission coefficients of different types of pollutants corresponding to diesel generators and distribution networks.

[0087] While solving the operating cost of the system established above, many constraints must be met. The specific constraints are as follows:

[0088] 1) Supply and demand constraints

[0089]

[0090] In the formula, Indicates the demand size of microgrid m at time t, including basic load and electric vehicle load

[0091] 2) Power Constraints

[0092]

[0093] In the formula, k represents different microgrid units, specifically and Indicates the upper and lower limits of the charging and discharging power of the micro power source, power interaction and energy storage system, and electric vehicle battery respectively.

[0094] 3) Climbing constraints

[0095]

[0096] In the formula, and They represent the DE output at time t and time t-1 in the scheduling period, R up(m) and R down(m) are the ramp-up and ramp-down rates of DE in each microgrid, respectively.

[0097] 4) Energy storage battery constraints

[0098]

[0099] In the formula, is the state of charge of the energy storage battery m at time t in the microgrid, and the upper and lower limits of the state of charge of the energy storage battery are and and They are the total power consumption of microgrid m energy storage dispatch.

[0100] Step S3, a vulture algorithm cluster system based on multi-hybrid strategy and adaptive variable improvement is proposed to solve the scheduling model and obtain the optimal coordinated power and micro-source output solution. The main improvements are as follows:

[0101] S31. Population initialization

[0102] Specifically, through the initial feasible solution X=(x1,x2,...,x d ),use Construct the reverse solution and determine the new combined population The best individuals serve as the new generation of elite individuals The elite reverse solution generation formula is used to determine a new individual population, and this population is used as the initial population.

[0103]

[0104] Where, K∈rand(), α j and β j It is a dynamic boundary.

[0105] S32. Improvements in the spatial selection phase

[0106] LV flight complies with the power law formula L(s)~|s| -1-β , the step size formula is defined as follows, and Alpha is dynamically adjusted by the algorithm iteration situation.

[0107]

[0108] In the early stage of optimization, a larger random step size is selected, and the algorithm gets rid of the interval where the local optimal solution is located. In the later stage, a smaller random step size is selected to search in the area near the optimal solution. The step size adaptive coefficient is defined as follows:

[0109] Alpha = w min +(w max -w min ) / 2*(1+cos(π*(iter-1) / (ger-1)))

[0110] In the formula, w min Take 5, w max Take 10, iter is the current iteration number, ger is the algorithm iteration number.

[0111] At this stage, although the individual update uses the average information of all individuals and the optimal individual information, the control variable parameter α and the random parameter r are prone to fall into the local optimum, so the LV flight strategy is introduced for updating. The corresponding spatial position update formula is:

[0112]

[0113] Set the state threshold trail. If the global optimal fitness after the algorithm population is updated is worse than the global optimal fitness of the previous generation, the threshold trail+1. If the global optimal solution is better than the optimal solution of the previous iteration, the threshold trail=0. When updating trail during the optimization process, if it is found that the limit is exceeded, the LV flight strategy update formula is used for space selection, otherwise the BES update formula is used for selection.

[0114] S33. Dive phase improvements

[0115] After determining the best point in the dive phase, the golden sine strategy is used to achieve a balance between local and global searches using its strong global search capability. The introduced golden section coefficient gradually narrows the search space and strengthens the search in the local space. The position transformation formula is:

[0116]

[0117] In the formula, represents the spatial position of the i-th individual in the t-th iteration process, P i t =(P i,1 ,P i,2 ,...,P i,d ) represents the optimal position of the i-th individual after t iterations. R1∈[0,2π] is a random number that determines the individual moving distance in the next iteration, and R2∈[0,π] is a random number that determines the direction of the next movement. x1=-π+(1-τ)*2π、x2=-π+τ*2π are based on the golden ratio coefficients The resulting irrational number.

[0118] S34.T distribution dynamic disturbance

[0119] T distribution disturbance is added to the global optimal individual, and the variable factor is adaptively updated in combination with the iteration situation to ensure the global search capability in the early stage of iteration and give full play to the potential optimal solution mining capability of the algorithm in the later stage. The form is P new,best =P best *(1+trnd(iter)), so that the algorithm searches near the global optimum after each iteration, effectively improving the convergence speed of the algorithm.

[0120] Based on the above improvements, the algorithm flow based on multi-hybrid strategy and adaptive variable improvement proposed in this paper is as follows: Figure 2 shown.

[0121] Step S4, the economic C obtained by the algorithm in S3 ec and environment C en The optimal solutions are denoted as C ec_min and C en_min , and the comprehensive cost target, construct a single-objective optimization function to evaluate the sub-goal satisfaction with the fuzzy membership function:

[0122] Sat=min(min(exp((C ec_min -C ec ) / C ec_min ),exp((C en_min -C en ) / C en_min )),1)

[0123] Example verification:

[0124] Taking the actual data of a digital breeding park as an example, the energy storage battery, diesel generator and photovoltaic capacity and parameters are set according to the actual needs of breeding scale, economic conditions and green development. At the same time, the photovoltaic and electricity load data of a typical day are as follows: Figure 3 The power parameters and pollutant emission coefficients of the microgrid components are shown in Table 1-2:

[0125] Table 1 Microgrid parameters

[0126]

[0127] Table 2 Pollutant emission coefficients

[0128]

[0129] Based on the above parameters, this paper uses MATLAB software to solve the optimization models in different scenarios for the superiority of the improved vulture algorithm for multi-objective optimization of cluster systems. The improved vulture algorithm sets the position change parameter a to 2, the search angle parameter a1 to 10, the search cycle R to 1.5, and the vulture movement intensity random numbers c1 and c2 to 1.8. For fair and just performance comparison, all other algorithms used are set to a population size of 200 and a number of iterations of 400. The scheduling step is 1h and the scheduling cycle is 24h. Set the following two optimization scenarios:

[0130] Scenario 1: Linearly weight the environmental and economic objective functions in the scheduling plan, comprehensively consider the economic and environmental goals, and verify the characteristics of different algorithms.

[0131] Scenario 2: Use the fuzzy membership function to comprehensively evaluate the optimization satisfaction of each sub-goal and reconstruct the objective function for optimization.

[0132] Simulation results:

[0133] 1) Analysis of optimization results of different algorithms

[0134] Table 3 gives a quantitative comparison of the scheduling results, specifically listing the optimal comprehensive operating cost and the optimization results of each sub-goal under the scheduling model of scenario 1. Compared with the traditional BES algorithm, the optimal cost of the improved BES algorithm is 51.9%, 44% and 62.6% lower in economic cost, environmental cost and comprehensive cost respectively. At the same time, the cost of the improved BES algorithm is also the best in the optimization results of each sub-goal. Although the LV-IBES algorithm has reduced the environmental cost and economic cost by 51.9% and 36.6% respectively, and the comprehensive cost by 47.8% compared with the BES algorithm, it still has not gotten rid of the local optimal problem. The overall performance of the algorithm in this paper is the best.

[0135] Table 3 Microgrid comprehensive cost optimization results

[0136]

[0137] The convergence curves of various algorithms are as follows: Figure 4 As shown in the figure, the improved BES algorithm has the highest convergence accuracy. The LV-IBES algorithm and the PSO algorithm converge at about 270 and 200 generations, respectively, while the improved BES algorithm converges at about 250 generations. Although the convergence speed is slower than that of PSO, the convergence accuracy is significantly improved, and the convergence accuracy is the highest in comparison. This shows that the algorithm proposed in this paper has achieved good results in both accuracy and speed. This shows that the proposed algorithm can more effectively coordinate the power interaction between microgrids, formulate internal power generation plans, and improve the economy of system operation.

[0138] 2) Analysis of operating conditions of micro-sources in microgrids under different scenarios

[0139] The results of different optimization scenarios of each microgrid are as follows: Figure 6 As shown. Generally speaking, compared with the purchase of electricity from the power grid, the output of diesel generators brings greater pollution costs. Therefore, the diesel generators in scenario 2 appropriately increase the power generation time compared with scenario 1, and the output power fluctuation does not change much compared with scenario 1. Compared with scenario 1, in scenario 2, the time for microgrid MG1 to sell electricity to the power grid increases from the original 5 hours to 9 hours, and the time for microgrid MG2 to sell electricity increases from 7 hours to 9 hours. At the same time, considering the characteristics of energy storage with strong pollution-free flexibility, the discharge depth and charge and discharge times are appropriately increased in scenario 2. The load demand capacity of microgrid MG3 is large, and its demand changes are relatively stable. It is also in the peak load period. In scenario 2, the power purchase from the large power grid is increased, and the output power of DE is appropriately adjusted. At the same time, the redundant power from MG1 and MG2 is more effectively utilized. Therefore, the optimization scenario using fuzzy membership satisfaction can prevent environmental pollution problems caused by relying on the large power grid and diesel power generation. At the same time, the complementary power between microgrids is further utilized to improve the flexibility of system operation.

[0140] 3) Analysis of optimization results in different scenarios

[0141] It can be seen from Table 4 that the single-objective optimization problem can make the optimization sub-objective reach the global optimal solution, but the optimization result of the corresponding sub-objective is difficult to achieve the ideal result, and the satisfaction evaluation results are low, which are 0.9088 and 0.8739 respectively. The comprehensive cost optimization takes into account the optimization results of economy and environment to a certain extent, but different weight values ​​often have different optimization results. Reasonable values ​​can improve the satisfaction of the objective function, but most results have strong uncertainty. The economic and environmental satisfactions of the linear weight comprehensive objective optimization results are 0.9289 and 0.9373 respectively, and the optimization results are relatively poor. The fuzzy membership evaluation satisfaction is selected to reconstruct the optimization target, and its economic, environmental and comprehensive costs are reduced by 390 yuan, 163 yuan and 553 yuan respectively relative to the linear cost, and the sub-objective satisfaction is improved to 0.9883 and 0.9515. The comparison of the optimization results under the two scenarios of the improved algorithm is shown in Table 4 below:

[0142] Table 4 Fuzzy optimization results

[0143]

[0144] The interactive power of microgrid tie lines after fuzzy membership optimization is as follows: Figure 5 As shown. In most periods of the dispatching cycle, the interconnection line interaction power is large. In the period of 00:00-05:00, the photovoltaic output is 0. The principle of "more output and less input" is used to achieve power coordination between microgrids, which can reduce the output power of diesel generators in the comprehensive linear optimization scenario, reduce environmental pollution and operating costs. In the peak photovoltaic output stage of 9:00-16:00, the photovoltaic output is sufficient to meet the load power demand. Under no special circumstances, the interconnection line interaction power is reduced, and the excess power is used to charge the energy storage, which is convenient for power consumption during the peak period of evening power consumption. Therefore, during the peak power consumption period of 19:00-24:00, multiple micro-power sources work together, and ES effectively releases the energy stored during the day during this period, resulting in a decrease in interaction power relative to the early power consumption valley period, and thus a coordination plan for the entire dispatching cycle is formulated.

[0145] The above are only preferred implementation modes of the present invention, which certainly cannot be used to limit the scope of rights of the present invention. It should be pointed out that for technicians in this technical field, any modification or equivalent replacement of the technical solution of the present invention without creative work does not depart from the protection scope of the technical solution of the present invention.

Claims

1. A multi-objective scheduling method for microgrid clusters based on an adaptive multi-hybrid strategy vulture algorithm, characterized in that: The following steps are involved: Step S1: obtaining the day-ahead forecast output data of the photovoltaic units of the multi-microgrid system, and obtaining the day-ahead forecast data of the load within the system according to the internal load characteristics of the microgrid; Step S2: Under the premise of stable operation of distributed power sources and installed capacity as constraints, a multi-objective scheduling model of economic and environmental costs is constructed by considering the power interaction between multiple microgrids; Step S3: improving the traditional vulture optimization algorithm based on adaptive variables combined with multi-optimization strategies, and using the improved algorithm to solve the microgrid cluster system scheduling model to obtain the optimal coordinated power and micro-source output solution; Step S4: Based on the established environmental and economic multi-objective model, the economic and environmental objectives are remodeled using fuzzy membership satisfaction to construct a new fitness function, and a scheduling plan that can take into account both economic and environmental optimization is obtained. A comparative analysis is then conducted between the proposed scheduling strategy and the traditional scheduling strategy to evaluate the effectiveness of the optimization method.

2. According to claim 1, a multi-objective scheduling method for microgrid clusters based on an adaptive multi-hybrid strategy vulture algorithm is characterized in that: In step S1, the day-ahead forecast data is acquired on a time scale of 1 hour, and the photovoltaic output and load of each microgrid system in 24 hours are 24 data points each, for a total of 144 data points.

3. According to claim 1, a multi-objective scheduling method for microgrid clusters based on an adaptive multi-hybrid strategy vulture algorithm is characterized in that: In step S2, the economic objective function 1 in the multi-objective optimization scheduling model of the cluster system is: In the formula, the economic cost includes the electric energy interaction Photovoltaic subsidies Micro power generation And equipment maintenance The costs incurred, Environmental objective function 2: Comprehensively consider comprehensive gas emissions to minimize environmental pollution In the formula, C k represents the cost of treating different types of pollutants, γ DE.k and γ GRID.k They are the emission coefficients of different types of pollutants corresponding to diesel generators and distribution networks.

4. According to claim 3, a multi-objective scheduling method for microgrid clusters based on an adaptive multi-hybrid strategy vulture algorithm is characterized in that: When solving the scheduling model, the power balance, output power, and capacity limit constraints are as follows: 1) Supply and demand constraints In the formula, Indicates the demand size of microgrid m at time t, including basic load and electric vehicle load 2) Power Constraints In the formula, k represents different microgrid units, specifically and Indicates the upper and lower limits of the charging and discharging power of the micro power source, power interaction and energy storage system, and electric vehicle battery respectively; 3) Climbing constraints In the formula, and They represent the DE output at time t and time t-1 in the scheduling period, R up(m) and R down(m) are the ramp-up and ramp-down rates of DE in each microgrid, respectively; 4) Energy storage battery constraints In the formula, is the state of charge of the energy storage battery m at time t in the microgrid, and the upper and lower limits of the state of charge of the energy storage battery are and They are the total power consumption of microgrid m energy storage dispatch.

5. According to claim 1, a multi-objective scheduling method for microgrid clusters based on an adaptive multi-hybrid strategy vulture algorithm is characterized in that: In step S3, in order to meet the needs of solving the optimization scheduling model, the BES algorithm performs the following improvement process: S31. Population initialization; S32. Improvement of spatial selection stage; S33. Dive phase optimization; S34.T distribution dynamic perturbations.

6. The multi-objective scheduling method of microgrid cluster based on adaptive multi-hybrid strategy vulture algorithm according to claim 5 is characterized in that: The S31 is specifically: using Constructing the inverse solution The best individual in the combined population is used as the new generation of elite individuals Determine the new individual population through the elite reverse solution generation formula: Where, K∈rand(), α j and β j It is a dynamic boundary.

7. The multi-objective scheduling method of microgrid cluster based on adaptive multi-hybrid strategy vulture algorithm according to claim 5 is characterized in that: The S32 is specifically as follows: select the LV strategy combined with the adaptive dynamic factor, define the step length formula as follows, and Alpha is dynamically adjusted by the algorithm iteration situation. In the early stage of optimization, a larger random step size is selected so that the algorithm can get rid of the interval where the local optimal solution is located. In the later stage, a smaller random step size is selected to search in the area near the optimal solution. The adaptation coefficient is as follows: Alpha=w min +(in max -In min ) / 2*(1+cos(π*(iter-1) / (ger-1))) In the formula, w min Take 5, w max Take 10, iter is the current iteration number, ger is the algorithm iteration number, The control variable parameter α and the random parameter r are prone to fall into local optimality. The LV flight strategy is introduced for updating, and the corresponding spatial position update formula is selected: Set the threshold trail. If the global optimal fitness after the algorithm population is updated is worse than the global optimal fitness of the previous generation, the threshold trail+1 is used. If the global optimal solution is better than the optimal solution of the previous iteration, the threshold trail=0. If the limit is exceeded when updating the trail during the optimization process, the LV update formula is used for space selection, otherwise the BES update formula is used for selection.

8. The multi-objective scheduling method of microgrid cluster based on adaptive multi-hybrid strategy vulture algorithm according to claim 5 is characterized in that: The S33 is specifically: using the strong global search capability of the golden sine strategy to achieve a balance between local and global searches, introducing the golden section coefficient to gradually narrow the search space, and strengthen the search of the local space. The position transformation formula is: In the formula, represents the spatial position of the i-th individual at the t-th iteration, represents the optimal position of the i-th individual after t iterations, R1∈[0,2π] is a random number that determines the individual moving distance in the next iteration, R2∈[0,π] is a random number that determines the direction of the next movement, x1=-π+(1-τ)*2π, x2=-π+τ*2π are based on the golden ratio coefficient The resulting irrational number.

9. The multi-objective scheduling method of microgrid cluster based on adaptive multi-hybrid strategy vulture algorithm according to claim 5 is characterized in that: The S34 is specifically as follows: adding T distribution disturbance to the global optimal individual, and adaptively updating the variable factor in combination with the iteration situation, to ensure the global search capability in the early stage of iteration, and to give full play to the potential optimal solution mining capability of the algorithm in the later stage, in the form of P new,best =P best *(1+trnd(iter)).

10. The multi-objective scheduling method of microgrid cluster based on adaptive multi-hybrid strategy vulture algorithm according to claim 1, characterized in that: In step S4, considering the satisfaction of each sub-goal optimization goal, combined with the multi-objective model of step S2, the optimization objective function is reconstructed as follows: Sat=min(min(exp((C ec_min -C ec ) / C ec_min ),exp((C en_min -C en ) / C en_min )),1)。

Images