Energy-carbon collaborative source-load storage optimization method based on improved multi-objective optimization algorithm

Abstract

The invention discloses an energy-carbon collaborative source-load storage optimization method based on an improved multi-objective optimization algorithm. Obtaining historical data of new energy output and load demand to generate a typical day scene; based on a typical day scene, constructing an energy-carbon collaborative source-load-storage optimization planning model, and solving the model by using three objective functions to obtain a minimum value of each objective function; setting the maximum value of each objective function, and constructing a Pareto plane; selecting a group of reference points on the Pareto plane, and performing boundary search along the normal vector direction of the Pareto plane to obtain a group of candidate non-dominated solutions; performing uniformity judgment on the candidate non-dominated solution; if the candidate non-dominated solution is not uniform, re-selecting a reference point; and if the candidate non-dominated solutions are uniform, outputting the candidate non-dominated solutions as a final Pareto optimal solution set. According to the method, the uniformity of the solution in the Pareto frontier can be improved through self-adaptive adjustment of the uniformity of the point set, so that a more effective solution strategy and technical support are provided for an energy-carbon collaborative source-load-storage optimization planning model.

Description

Technical Field

[0001] This invention relates to the field of nano-micro power grid planning, and in particular to an energy-carbon synergistic source-load-storage optimization method based on an improved multi-objective optimization algorithm. Background Technology

[0002] Microgrids, as an important carrier for achieving a green and low-carbon energy transition, are being widely promoted and applied in distributed energy systems. Furthermore, nanogrids, as smaller microgrids with more concentrated geographical locations, offer greater operational flexibility on a smaller spatial scale. Microgrids and nanogrids together form nano-microgrid clusters, integrating diverse energy supply methods and load characteristics, and incorporating energy storage devices to effectively improve the operational resilience and reliability of the energy system. In the source-load-storage planning models of nano-microgrid clusters, in addition to economic efficiency as the objective function, carbon emission factors and operational stability factors in the energy system are often considered, thus balancing multi-scenario and multi-objective needs.

[0003] However, in the field of nano-microgrid planning, a pressing technical problem is how to solve the constructed energy-carbon co-source-load-storage optimization planning model using an improved multi-objective optimization algorithm to enhance the uniformity of the solution set distribution on the Pareto front. Existing multi-objective optimization models often employ standard multi-objective particle swarm optimization algorithms and multi-objective genetic algorithms, but these often suffer from problems such as getting trapped in local optima and uneven solution set distribution when faced with large-scale, strongly coupled variables, severely limiting the model's solution performance and practicality. Therefore, designing an improved multi-objective optimization algorithm and effectively applying it to the energy-carbon co-source-load-storage optimization planning model to improve the uniformity of the solution set distribution is of significant research value. Summary of the Invention

[0004] Based on the above background, this invention proposes an energy-carbon synergistic source-load-storage optimization planning model based on an improved multi-objective optimization algorithm. At the algorithm level, the existing multi-objective optimization solution strategy is improved to enhance the uniformity of the solution on the Pareto front, thereby providing a more effective solution strategy and technical support for the planning of nano-micro power grid groups.

[0005] The technical solution adopted in this invention is: S1. Obtain historical data on new energy output and load demand, and then generate typical daily scenarios through kernel density estimation and correlation processing; S2. Construct an energy-carbon co-source-load-storage optimization planning model based on typical daily scenarios, and solve the energy-carbon co-source-load-storage optimization planning model with three preset objective functions to obtain the minimum value of each objective function. S3. Set the maximum value according to the minimum value of each objective function, and normalize each objective function according to the minimum and maximum values ​​of each objective function, and then construct the Pareto plane according to the normalization results. S4. Select a set of uniformly distributed reference points along a straight line on the Pareto plane, and then perform a boundary search along the direction of the normal vector of the Pareto plane that reduces the objective function, starting from each reference point, to obtain a set of candidate non-dominated solutions. S5. Determine the uniformity of the candidate non-dominated solutions: If they are not uniform, return to step S4 and reselect the reference point; if they are uniform, output the candidate non-dominated solutions as the final Pareto optimal solution set.

[0006] Step S1 specifically involves: S1.1 Obtain historical data on renewable energy output and load demand, and then use kernel density estimation to construct the probability density functions of renewable energy output and load demand respectively; S1.2. Based on the probability density functions of new energy output and load demand, the Frank Copula function is used to process the correlation and construct a joint probability distribution function. S1.3. Randomly sample the joint probability distribution function multiple times to obtain several sampled values, and then perform inverse transformation on the several sampled values ​​according to the joint probability distribution function to obtain the original daily scene set; S1.4. Use the K-means clustering algorithm to reduce the original daily scene set to obtain typical daily scenes.

[0007] Step S3 specifically involves: S3.1 Set a preset multiplier, multiply the minimum value of each objective function by the preset multiplier to obtain the maximum value of each objective function; S3.2. Normalize each objective function based on its minimum and maximum values, and then construct a three-dimensional coordinate system based on the values ​​of each objective function. S3.3 In the three-dimensional coordinate system, obtain the single-objective optimization boundary points based on the minimum and maximum values ​​of each objective function; S3.4 The Pareto plane is obtained by fitting the boundary points of the single-objective optimization using the least squares method.

[0008] In step S2, the three preset objective functions are constructed with minimizing the total planning and operating cost, minimizing carbon emissions, and minimizing load shedding as optimization objectives, respectively. The single-objective optimization boundary points are set according to the following formula: in, This represents the single-objective optimization boundary point that minimizes the total planning and operating cost. This represents the objective function value for carbon emissions that minimizes the total planning and operating costs. This represents the objective function value of the load shedding amount that minimizes the total planning and operating costs. This represents the single-objective optimization boundary point that maximizes the total planning and operating cost. This represents the objective function value for carbon emissions that maximizes the total planning and operating costs. This represents the objective function value of the load shedding amount that maximizes the total planning and operating costs. This represents the single-objective optimization boundary point that minimizes carbon emissions. This represents the total planning and operating cost objective function value that minimizes carbon emissions. This represents the objective function value of the load shedding amount that minimizes carbon emissions. This represents the single-objective optimization boundary point that maximizes carbon emissions. This represents the objective function value of the total planning and operating cost that maximizes carbon emissions. This represents the objective function value of the load shedding amount that maximizes carbon emissions; This represents the single-objective optimization boundary point that minimizes the load shedding amount. This represents the objective function value of the total planning and operating cost that minimizes the load shedding. This represents the objective function value for carbon emissions that minimizes the load shedding rate. This represents the single-objective optimization boundary point that minimizes the load shedding amount. This represents the objective function value of the total planning and operating cost that minimizes the load shedding. This represents the objective function value for carbon emissions that minimizes the load shedding.

[0009] In step S4, the uniformly distributed reference points specifically refer to the equal L2 distance between reference points of adjacent sequences. The adjacent sequence refers to the sequence formed by the shortest pairwise Euclidean distance between reference points.

[0010] Step S5 specifically involves: S5.1 Map a set of candidate non-dominated solutions to a three-dimensional coordinate system to obtain the coordinate points corresponding to each candidate non-dominated solution; S5.2 Calculate the L2 distance and the average distance between the coordinate points corresponding to the candidate non-dominated solutions of adjacent sequences using the following formula: in, This represents the L2 distance between the coordinates of candidate non-dominated solutions in adjacent sequences. This represents the average distance. Represents the coordinates of a point set. This indicates the calculation of the L2 norm. This represents the total number of non-dominated solutions; S5.3, Perform uniformity judgment on candidate non-dominated solutions: If the L2 distance between the coordinate points corresponding to the candidate non-dominated solutions of adjacent sequences is not within the preset distance range, then the reference point is reselected according to the following formula; in, Indicates the three-dimensional position of the coordinate point. This indicates the selection of a new reference point. Indicates the original reference point. Indicates the scaling factor; If the L2 distance between the coordinates of the candidate non-dominated solutions in adjacent sequences is within a preset distance range, then the candidate non-dominated solutions will be output as the final Pareto optimal solution set.

[0011] In step S2, the three preset objective functions are constructed with minimizing the total planning and operating cost, minimizing carbon emissions, and minimizing load shedding as optimization objectives, respectively. (1) The objective function is constructed with minimizing the total planning and operating cost as the optimization objective, and is specifically set according to the following formula: In the formula, Let the objective function be the total planning and operating cost. To cover expansion costs, For operating costs, For the number of scenes, For the number of time periods, In order to plan the equipment collection, Cost of expanding unit equipment capacity To expand the capacity of the equipment, For the collection of operating equipment, Indicates the first Unit power operating cost coefficient of each device Representing a scene ,time The The power of each device; (2) The objective function, which is constructed with minimizing carbon emissions as the optimization goal, is set according to the following formula: In the formula, The objective function is constructed with the goal of minimizing carbon emissions. As a carbon emission factor, For the scene ,time The carbon emission power of diesel generators; (3) The objective function is constructed with minimizing the load shedding amount as the optimization objective, and is specifically set according to the following formula: In the formula, The objective function is constructed with the goal of minimizing the load shedding amount. Representing a scene ,time The amount of power cut-off load, Representing a scene ,time heat load, Representing a scene ,time The amount of cooling load, Indicates electrical load. Indicates heat load, This indicates the cooling load.

[0012] The energy-carbon synergistic source-load-storage optimization planning model in step S2 includes external power supply constraints, diesel generator constraints, ground source heat pump constraints, energy storage battery constraints, flexible load constraints, and network balance constraints.

[0013] The beneficial effects of this invention are: This invention proposes an energy-carbon co-source-load-storage optimization planning model based on an improved multi-objective optimization algorithm. By employing an improved multi-objective optimization algorithm, the aim is to enhance the uniformity of solutions in the Pareto front, thereby providing a more effective solution strategy and technical support for the energy-carbon co-source-load-storage optimization planning model.

[0014] The constructed energy-carbon co-source-load-storage optimization planning model based on the improved multi-objective optimization algorithm can improve the uniformity of the solution in the Pareto front through adaptive adjustment of the point set uniformity, thus providing a more effective solution strategy and technical support for the energy-carbon co-source-load-storage optimization planning model. Attached Figure Description

[0015] Figure 1 This is a flowchart of the method in this embodiment; Figure 2 This is a schematic diagram of the Pareto plane in this embodiment; Figure 3 This is a schematic diagram of a reference point on the Pareto plane in this embodiment; Figure 4 This is a schematic diagram illustrating the search for candidate non-dominated solutions in this embodiment. Detailed Implementation

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and do not limit the scope of protection of this invention.

[0018] like Figure 1 As shown, this embodiment includes the following steps: S1. Obtain historical data on renewable energy output and load demand, and then generate typical daily scenarios of renewable energy-load through kernel density estimation and correlation processing; S2. Based on the constructed typical daily scenario of new energy-load, construct an energy-carbon synergistic source-load-storage optimization planning model, and solve the energy-carbon synergistic source-load-storage optimization planning model with three preset objective functions to obtain the minimum value of each objective function. S3. Set the maximum value according to the minimum value of each objective function, and normalize each objective function according to the minimum and maximum values ​​of each objective function, and then construct the Pareto plane according to the normalization results. S4, such as Figure 3 As shown, a set of uniformly distributed reference points are selected along a straight line on the Pareto plane. Then, a set of candidate non-dominated solutions is obtained by performing a boundary search along the direction of the normal vector of the Pareto plane that reduces the objective function, starting from each reference point. S5. Determine the uniformity of the candidate non-dominated solutions: If they are not uniform, return to step S4 and reselect the reference point; if they are uniform, output the candidate non-dominated solutions as the final Pareto optimal solution set.

[0019] Step S1 is as follows: S1.1 Obtain historical data on renewable energy output and load demand, and then use kernel density estimation to construct hourly probability density functions for renewable energy output and load demand respectively; Specifically, assuming random variables There are independent and identically distributed samples. Then point probability density function at as follows: In the formula: For kernel functions; The number of samples; For window width.

[0020] S1.2. Based on the hourly probability density functions of new energy output and load demand, the Frank Copula function is used to perform correlation processing to construct a joint probability distribution function; The Copula function is used to process the correlation. The general expression for Copula is: In the formula: for The joint distribution function of n random variables; Let be the marginal distribution function of the random variable; This is the Copula connection function.

[0021] This invention uses the Frank Copula function, whose expression is: In the formula: , It is a random variable; It is the parameter of the Frank Copula function.

[0022] S1.3. Randomly sample the joint probability distribution function multiple times to obtain several sampled values, and then perform inverse transformation on the several sampled values ​​according to the joint probability distribution function to obtain the original daily scene set; Furthermore, the generation of massive new energy-load scenarios involves the following steps: (1) Randomly sample the joint probability distribution function of each time period to obtain the sampled values.

[0023] (2) Based on the joint probability distribution function, the sampled values ​​are inversely transformed to obtain the output and load values ​​of new energy sources at each time period.

[0024] (3) Repeat steps (1) and (2) to obtain a large number of data considering the correlation between new energy sources and load. Among them, a daily scenario is a two-dimensional set composed of new energy output and load in a time series.

[0025] S1.4. The K-means clustering algorithm is used to reduce the original daily scene set to obtain typical daily scenes of new energy and load.

[0026] This invention employs the k-means++ algorithm to reduce the scene, and the specific steps are as follows: (1) For all generated new energy-load daily scenarios, each generated scenario (a two-dimensional set of new energy output and load in time series) is regarded as a sample point, and a sample point is randomly selected for initialization as the first cluster center.

[0027] (2) Iterate through the data and calculate the probability that each sample point is selected as a cluster center, and select the remaining cluster centers. The probability is inversely proportional to the square of the distance, and the specific calculation method is as follows: In the formula: Sample points L2 distance to the nearest cluster center; It represents the total number of sample points.

[0028] (3) Repeat steps (1) and (2) until a selection is made. Until the initial cluster centers are identified.

[0029] (4) Use the selected initial cluster centers as the starting point and apply the standard k-means algorithm for clustering.

[0030] In (3), the number of clusters must be specified. The value of is determined using the silhouette coefficient to obtain the optimal number of clusters. For sample points... Its profile coefficient is defined as: In the formula: Indicates the minimum average distance between clusters; This represents the average distance within a cluster. By combining the density of samples within a cluster and the separation of samples between clusters, the clustering effect can be comprehensively evaluated.

[0031] For the entire sample, the silhouette coefficient is defined as the average of the silhouette coefficients of all sample points: In summary, the k-means++ clustering algorithm is used as the scene reduction method, and the optimal number of clusters is determined by the silhouette coefficient. Finally, the desired number of clusters is generated. There are several clusters, and the center of each cluster is the curve of a typical daily scene. The proportion of sample points contained in each cluster is used as the weight value of that scene.

[0032] By using the above-mentioned scene generation and scene reduction methods, multiple scenarios suitable for extended planning are obtained.

[0033] Step S3 is as follows: S3.1 Set a preset multiplier, multiply the minimum value of each objective function by the preset multiplier to obtain the maximum value of each objective function; S3.2. Normalize each objective function based on its minimum and maximum values, and then construct a three-dimensional coordinate system based on the values ​​of each objective function. S3.3 In the three-dimensional coordinate system, obtain the single-objective optimization boundary points based on the minimum and maximum values ​​of each objective function; S3.4 The Pareto plane is obtained by fitting the boundary points of the single-objective optimization using the least squares method, such as... Figure 2 As shown.

[0034] Since the maximum value of each objective function cannot be directly obtained through the optimization model, it is set as the minimum value of the objective function. times( Generally, 1000 can be used), and the process is as follows: After obtaining the maximum and minimum values, the three objective functions are subjected to maxima-minus normalization so that the objective function values ​​are all between [0,1]. Accordingly, after maxima-minima normalization, we have , .

[0035] After calculating the maximum and minimum values ​​of each objective function, six distinct points in the plane can be obtained. The following are, in order: After normalization by the maximum and minimum, it becomes .

[0036] Single-objective optimization boundary points are set according to the following formula: in, This represents the single-objective optimization boundary point that minimizes the total planning and operating cost; This represents the objective function value for carbon emissions that minimizes the total planning and operating costs. This represents the objective function value of the load shedding amount that minimizes the total planning and operating costs. This represents the single-objective optimization boundary point that maximizes the total planning and operating cost; This represents the objective function value for carbon emissions that maximizes the total planning and operating costs. This represents the objective function value of the load shedding amount that maximizes the total planning and operating costs. This represents the single-objective optimization boundary point that minimizes carbon emissions. This represents the objective function value of the total planning and operating cost that minimizes carbon emissions. This represents the objective function value of the load shedding amount that minimizes carbon emissions; This represents the single-objective optimization boundary point that maximizes carbon emissions; This represents the objective function value of the total planning and operating cost that maximizes carbon emissions. This represents the objective function value of the load shedding amount that maximizes carbon emissions; This represents the single-objective optimization boundary point that minimizes the load shedding amount; This represents the objective function value of the total planning and operating cost that minimizes the load shedding amount; This represents the objective function value for carbon emissions that minimizes the load shedding rate; This represents the single-objective optimization boundary point that minimizes the load shedding amount; This represents the objective function value of the total planning and operating cost that minimizes the load shedding amount; This represents the objective function value for carbon emissions that minimizes the load shedding.

[0037] In step S4, the uniformly distributed reference points are specifically defined as reference points of adjacent sequences having equal L2 distances. The adjacent sequence refers to the sequence formed by the shortest pairwise Euclidean distance between reference points.

[0038] In step S4, a set of candidate non-dominated solutions is obtained as follows: First, a set of uniformly distributed reference points are selected along a straight line on the Pareto plane; second, starting from each reference point, feasible solutions are searched along the normal vector direction of the Pareto plane that reduces the objective function. The farthest feasible solution along this direction is the candidate non-dominated solution for each reference point (solutions further away are infeasible, i.e., at least one equality constraint or inequality constraint is violated). The candidate non-dominated solutions of each reference point constitute the set of candidate non-dominated solutions. A non-dominated solution is one in the feasible solution space that is superior to any other solution in all objective functions.

[0039] That is, passing through each point By drawing a perpendicular line to the Pareto plane, we can obtain the set of points projected onto the Pareto front. This yields the Pareto front point set. The projection process is as follows: Figure 4As shown, the solid line represents the Pareto front plane, and the dashed line represents the Pareto front surface.

[0040] like Figure 4 As shown, step S5 specifically involves: S5.1 Map a set of candidate non-dominated solutions to the three-dimensional coordinate system constructed in step S3.2 to obtain the coordinate points corresponding to each candidate non-dominated solution; S5.2 Calculate the L2 distance between the coordinate points corresponding to the candidate non-dominated solutions of adjacent sequences and the average distance between the coordinate points corresponding to the candidate non-dominated solutions of all adjacent sequences using the following formulas: in, This represents the L2 distance between the coordinates of candidate non-dominated solutions in adjacent sequences. This represents the average distance. Represents the coordinates of a point set. This indicates the calculation of the L2 norm. This represents the total number of non-dominated solutions (point sets); S5.3, Perform uniformity judgment on candidate non-dominated solutions: If the L2 distance between the coordinate points corresponding to the candidate non-dominated solutions of adjacent sequences is not within the preset distance range, then the reference point is reselected according to the following formula; in, Indicates the three-dimensional position of the coordinate point; This indicates the selection of a new reference point; Indicates the original reference point; This represents the scaling factor.

[0041] If the L2 distance between the coordinates of the candidate non-dominated solutions in adjacent sequences is within a preset distance range, then the candidate non-dominated solutions will be output as the final Pareto optimal solution set.

[0042] Specifically, the first step is to calculate the L2 distance between labeled adjacent point sets on the Pareto front. and its average value .

[0043] The second step is to compare the sizes: Compare and , The size of can be any of the following three cases: (1) (2) (3) Step 3, Location Update: On the Pareto front plane, point The coordinates are updated as follows: In the formula: The parameter has a value range between [-1, 1].

[0044] After the location is updated By drawing a perpendicular line to the auxiliary surface of the Pareto front, we can obtain the set of points projected onto the Pareto front. Recalculate at this point. : (1) If Then continue to increase appropriately. ; (2) If Then continue to reduce appropriately. ; (3) If If not, then no update will be performed.

[0045] In summary, by updating the Pareto front surface point set, the Pareto front surface point set becomes more uniform, thus providing a more effective solution strategy and technical support for the energy-carbon co-source-load-storage optimization planning model.

[0046] In step S2, the three preset objective functions are constructed with minimizing the total planning and operating cost, minimizing carbon emissions, and minimizing load shedding as optimization objectives, respectively. (1) The objective function is constructed with minimizing the total planning and operating cost as the optimization objective, and is specifically set according to the following formula: In the formula, Let the objective function be the total cost. To cover expansion costs, For operating costs, For the number of scenes, For the number of time periods, The planned equipment suite includes energy storage batteries, flexible loads, and diesel generators. Cost of expanding unit equipment capacity To expand the capacity of the equipment, The system comprises a collection of operating equipment, including five types: external power supply, diesel generators, ground source heat pumps, energy storage batteries, and flexible loads. Indicates the first Unit power operating cost coefficient of each device Representing a scene ,time The The power of each device; (2) The objective function, which is constructed with minimizing carbon emissions as the optimization goal, is set according to the following formula: In the formula, The objective function is an environmentally friendly one, meaning it is constructed with the goal of minimizing carbon emissions. As a carbon emission factor, For the scene ,time The carbon emission power of diesel generators; (3) The objective function is constructed with minimizing the load shedding amount as the optimization objective, and is specifically set according to the following formula: In the formula, The resilience enhancement objective function is a function constructed with minimizing the load shedding amount as the optimization objective. Representing a scene ,time The amount of power cut-off load, Representing a scene ,time heat load, Representing a scene ,time The cooling load, where the variable superscript Indicates electrical load. Indicates heat load, Indicates cooling load; The energy-carbon synergistic source-load-storage optimization planning model in step S2 includes external power supply constraints, diesel generator constraints, ground source heat pump constraints, energy storage battery constraints, flexible load constraints, and network balance constraints. 1) External power supply constraints are set according to the following formula: In the formula, For the scene ,time External power supply power, This is the upper limit of the external power supply. 2) Diesel generator constraints are set according to the following formula: In the formula, For the scene ,time The output power of the diesel generator, Maximum active power of diesel generators This is the lower limit of the active power of a diesel generator. This represents the maximum uphill climbing power of the diesel generator. This refers to the maximum downhill climbing power of the diesel generator; 3) The ground source heat pump's constraint regulation capability mainly depends on changes in ground source temperature and pump flow rate, and is set according to the following formula: In the formula, For the scene ,time The input electrical power of the ground source heat pump, The cooling energy efficiency ratio of a ground source heat pump. The heating efficiency ratio of a ground source heat pump. For the scene ,time The cooling power regulation capability of the ground source heat pump For the scene ,time The ground source heat pump's heating power regulation capability, For the scene ,time The ground source heat pump's cooling and heating status is a 0-1 variable. Indicates in the scene ,time The ground source heat pump is in heating mode. Indicates in the scene ,time The ground source heat pump is in cooling mode. , These are the upper and lower limits of the output heat power of the ground source heat pump, respectively. , These are the upper and lower limits of the cooling power output of the ground source heat pump, respectively. 4) Energy storage battery constraints include power balance constraints, charge / discharge power constraints, and energy storage capacity constraints, which are set according to the following formulas: In the formula, For the scene ,time The charging power of the energy storage battery, For the scene ,time The discharge power of the energy storage battery, This is the upper limit of the charging power of energy storage batteries. This represents the upper limit of the discharge power of energy storage batteries. For the scene ,time The state of charge of the energy storage battery is a 0-1 variable. For the scene ,time The state of discharge of the energy storage battery is a 0-1 variable. For the scene ,time The energy storage battery capacity, For the scene ,time The energy storage battery capacity, To improve the charging and discharging efficiency of energy storage batteries, , These are the upper and lower limits of the energy storage battery, respectively. 5) Flexible load constraints are set according to the following formula: In the formula, For the scene ,time Flexible load power, This represents the upper limit of the adjustable range for flexible loads. This represents the lower limit of the adjustable range for flexible loads.

[0047] 6) Network balance constraints are set according to the following formula: In the formula, For the scene ,time electrical load power, For the scene ,time Heat load power, For the scene ,time The cooling load power.

[0048] The above detailed embodiments illustrate the technical solution and beneficial effects of the present invention. It should be understood that the above description is only the most preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, additions, and equivalent substitutions made within the scope of the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for optimizing energy and carbon co-source load and storage based on an improved multi-objective optimization algorithm, characterized in that, The method includes the following steps: S1. Obtain historical data on new energy output and load demand, and then generate typical daily scenarios through kernel density estimation and correlation processing; S2. Construct an energy-carbon co-source-load-storage optimization planning model based on typical daily scenarios, and solve the energy-carbon co-source-load-storage optimization planning model with three preset objective functions to obtain the minimum value of each objective function. S3. Set the maximum value according to the minimum value of each objective function, and normalize each objective function according to the minimum and maximum values ​​of each objective function, and then construct the Pareto plane according to the normalization results. S4. Select a set of uniformly distributed reference points along a straight line on the Pareto plane, and then perform a boundary search along the direction of the normal vector of the Pareto plane that reduces the objective function, starting from each reference point, to obtain a set of candidate non-dominated solutions. S5. Determine the uniformity of the candidate non-dominated solutions: If they are not uniform, return to step S4 and reselect the reference point; if they are uniform, output the candidate non-dominated solutions as the final Pareto optimal solution set.

2. The energy-carbon synergistic source-load-storage optimization method based on an improved multi-objective optimization algorithm according to claim 1, characterized in that: Step S1 specifically involves: S1.1 Obtain historical data on renewable energy output and load demand, and then use kernel density estimation to construct the probability density functions of renewable energy output and load demand respectively; S1.

2. Based on the probability density functions of new energy output and load demand, the Frank Copula function is used to process the correlation and construct a joint probability distribution function. S1.

3. Randomly sample the joint probability distribution function multiple times to obtain several sampled values, and then perform inverse transformation on the several sampled values ​​according to the joint probability distribution function to obtain the original daily scene set; S1.

4. Use the K-means clustering algorithm to reduce the original daily scene set to obtain typical daily scenes.

3. The energy-carbon synergistic source-load-storage optimization method based on an improved multi-objective optimization algorithm according to claim 1, characterized in that: Step S3 specifically involves: S3.1 Set a preset multiplier, multiply the minimum value of each objective function by the preset multiplier to obtain the maximum value of each objective function; S3.

2. Normalize each objective function based on its minimum and maximum values, and then construct a three-dimensional coordinate system based on the values ​​of each objective function. S3.3 In the three-dimensional coordinate system, obtain the single-objective optimization boundary points based on the minimum and maximum values ​​of each objective function; S3.4 The Pareto plane is obtained by fitting the boundary points of the single-objective optimization using the least squares method.

4. The energy-carbon synergistic source-load-storage optimization method based on an improved multi-objective optimization algorithm according to claim 1, characterized in that: In step S2, the three preset objective functions are constructed with minimizing the total planning and operating cost, minimizing carbon emissions, and minimizing load shedding as optimization objectives, respectively. The single-objective optimization boundary points are set according to the following formula: in, This represents the single-objective optimization boundary point that minimizes the total planning and operating cost. This represents the objective function value for carbon emissions that minimizes the total planning and operating costs. This represents the objective function value of the load shedding amount that minimizes the total planning and operating costs. This represents the single-objective optimization boundary point that maximizes the total planning and operating cost. This represents the objective function value for carbon emissions that maximizes the total planning and operating costs. This represents the objective function value of the load shedding amount that maximizes the total planning and operating costs. This represents the single-objective optimization boundary point that minimizes carbon emissions. This represents the total planning and operating cost objective function value that minimizes carbon emissions. This represents the objective function value of the load shedding amount that minimizes carbon emissions. This represents the single-objective optimization boundary point that maximizes carbon emissions. This represents the objective function value of the total planning and operating cost that maximizes carbon emissions. This represents the objective function value of the load shedding amount that maximizes carbon emissions; This represents the single-objective optimization boundary point that minimizes the load shedding amount. This represents the objective function value of the total planning and operating cost that minimizes the load shedding. This represents the objective function value for carbon emissions that minimizes the load shedding rate. This represents the single-objective optimization boundary point that minimizes the load shedding amount. This represents the objective function value of the total planning and operating cost that minimizes the load shedding. This represents the objective function value for carbon emissions that minimizes the load shedding.

5. The energy-carbon synergistic source-load-storage optimization method based on an improved multi-objective optimization algorithm according to claim 1, characterized in that: In step S4, the uniformly distributed reference points specifically refer to the equal L2 distance between reference points of adjacent sequences. The adjacent sequence refers to the sequence formed by the shortest pairwise Euclidean distance between reference points.

6. The energy-carbon synergistic source-load-storage optimization method based on an improved multi-objective optimization algorithm according to claim 3, characterized in that: Step S5 specifically involves: S5.1 Map a set of candidate non-dominated solutions to a three-dimensional coordinate system to obtain the coordinate points corresponding to each candidate non-dominated solution; S5.2 Calculate the L2 distance and the average distance between the coordinate points corresponding to the candidate non-dominated solutions of adjacent sequences using the following formula: in, This represents the L2 distance between the coordinates of candidate non-dominated solutions in adjacent sequences. This represents the average distance. Represents the coordinates of a point set. This indicates the calculation of the L2 norm. This represents the total number of non-dominated solutions; S5.3, Perform uniformity judgment on candidate non-dominated solutions: If the L2 distance between the coordinate points corresponding to the candidate non-dominated solutions of adjacent sequences is not within the preset distance range, then the reference point is reselected according to the following formula; in, Indicates the three-dimensional position of the coordinate point. This indicates the selection of a new reference point. Indicates the original reference point. Indicates the scaling factor; If the L2 distance between the coordinates of the candidate non-dominated solutions in adjacent sequences is within a preset distance range, then the candidate non-dominated solutions will be output as the final Pareto optimal solution set.

7. The energy-carbon synergistic source-load-storage optimization method based on an improved multi-objective optimization algorithm according to claim 1, characterized in that: In step S2, the three preset objective functions are constructed with minimizing the total planning and operating cost, minimizing carbon emissions, and minimizing load shedding as optimization objectives, respectively. (1) The objective function is constructed with minimizing the total planning and operating cost as the optimization objective, and is specifically set according to the following formula: In the formula, Let the objective function be the total planning and operating cost. To cover expansion costs, For operating costs, For the number of scenes, For the number of time periods, In order to plan the equipment collection, Cost of expanding unit equipment capacity To expand the capacity of the equipment, For the collection of operating equipment, Indicates the first Unit power operating cost coefficient of each device Representing a scene ,time The The power of each device; (2) The objective function, which is constructed with minimizing carbon emissions as the optimization goal, is set according to the following formula: In the formula, The objective function is constructed with the goal of minimizing carbon emissions. As a carbon emission factor, For the scene ,time The carbon emission power of diesel generators; (3) The objective function is constructed with minimizing the load shedding amount as the optimization objective, and is specifically set according to the following formula: In the formula, The objective function is constructed with the goal of minimizing the load shedding amount. Representing a scene ,time The amount of power cut-off load, Representing a scene ,time heat load, Representing a scene ,time The amount of cooling load, Indicates electrical load. Indicates heat load, This indicates the cooling load.

8. The energy-carbon synergistic source-load-storage optimization method based on an improved multi-objective optimization algorithm according to claim 1, characterized in that: The energy-carbon synergistic source-load-storage optimization planning model in step S2 includes external power supply constraints, diesel generator constraints, ground source heat pump constraints, energy storage battery constraints, flexible load constraints, and network balance constraints.

Images