Electric heating energy system emission reduction cost evaluation method based on KPCA-AHC new energy dimension reduction clustering

Abstract

The invention discloses an electric heating energy system emission reduction cost evaluation method based on KPCA-AHC new energy dimension reduction clustering, and the method comprises the steps: firstly carrying out the dimension reduction of new energy data through employing a kernel principal component analysis method KPCA, and carrying out the output clustering of new energy through an AHC method based on the data after the dimension reduction, and obtaining a wind power and photovoltaic typical output scene; secondly, establishing a multi-target optimization scheduling model of the electric heating integrated energy system by taking the minimum operation cost and carbon emission as the target, and solving the model by adopting a normalized plane constraint method to obtain the Pareto leading edge of the system operation cost and carbon emission; and finally, evaluating the emission reduction cost of the system based on the system cost-carbon emission Pareto leading edge. Example results show that the method can better describe the uncertainty of new energy output, an optimal scheduling scheme of the system is obtained on the basis, the emission reduction cost of the system can be accurately measured and calculated, and reference is provided for evaluation of the emission reduction cost of the system.

Description

Technical Field

[0001] The present invention belongs to the field of low-carbon scheduling of integrated energy systems, and in particular to an emission reduction cost evaluation method for an electric-thermal integrated energy system based on KPCA-AHC new energy dimensionality reduction clustering. Background Art

[0002] As energy crisis, environmental pollution, climate change and other issues become increasingly prominent, giving priority to the development and consumption of new energy sources such as wind and light is the main way to reduce greenhouse gas emissions and reduce dependence on fossil energy. In order to achieve diversified energy demand, integrated energy systems have become an inevitable choice to promote energy transformation. When scheduling the electric-heat integrated energy system (EHIES) on the day before, it is necessary to take the predicted output of new energy as the planned output for scheduling. When the output forecast deviation is large, it will greatly affect the reliability of EHIES scheduling. Therefore, it is urgent to study the optimal scheduling of EHIES taking into account the uncertainty of new energy output. The emission reduction cost is defined as the potential cost of reducing unit carbon emissions. Too high emission reduction costs will hinder the decarbonization enthusiasm of high-carbon industries, and too low emission reduction costs will have an adverse impact on the carbon market. Therefore, accurate evaluation of the emission reduction cost of EHIES can provide reference and reference for its emission reduction strategy. Summary of the invention

[0003] Purpose of the invention: In view of the defects and shortcomings of the prior art, the present invention proposes a method for generating a set of typical output scenarios of new energy sources by integrating Kernel Principal Component Analysis (KPCA) and Agglomerative Hierarchical Clustering (AHC) technologies. First, the KPCA method is used to reduce the dimension of new energy data. Based on the reduced-dimensional data, the AHC technology is used to cluster the new energy output to obtain typical output scenarios of wind power and photovoltaic power. Secondly, with the goal of minimizing operating costs and carbon emissions, a multi-objective optimization scheduling model for an electric-thermal integrated energy system is established. The normalized plane constraint method is used to solve the model to obtain the system operating cost and carbon emission Pareto frontier. Finally, the emission reduction cost of the system is evaluated based on the system cost-carbon emission Pareto frontier. The kernel principal component dimensionality reduction clustering method can better characterize the uncertainty of new energy output, and on this basis, the optimal scheduling plan of the system can be obtained. The emission reduction cost of the system can be accurately calculated, providing a reference and reference for the emission reduction cost evaluation of the system.

[0004] Technical solution: The method for evaluating emission reduction costs of electric thermal energy systems based on KPCA-AHC new energy dimensionality reduction clustering of the present invention comprises the following steps:

[0005] Step 1: Collect new energy data (including new energy historical output data and meteorological data of the area where the new energy station is located), and use the kernel principal component analysis method KPCA to reduce the dimension of the new energy data to obtain new energy reduced dimension data;

[0006] Step 2: Based on the dimension reduction data of new energy, the output of new energy is clustered by using the agglomerative hierarchical clustering method AHC to obtain the typical output scenarios of wind power and photovoltaic power;

[0007] Step 3: With the goal of minimizing operating costs and carbon emissions, a multi-objective optimization scheduling model for the electric-thermal integrated energy system is established, and the normalized plane constraint method is used to solve the model to obtain the Pareto frontier of system operating costs and carbon emissions;

[0008] Step 4: Based on the obtained Pareto frontier, evaluate the carbon reduction cost of the electric thermal energy system.

[0009] Furthermore, step 1 is specifically as follows:

[0010] The new energy data includes wind power data and photovoltaic data;

[0011] The following is an example of wind power. Wind power output is not only strongly correlated with historical output, but also closely related to meteorological factors such as wind speed, wind direction, air pressure and temperature. The historical data composed of wind power output, wind speed, wind direction, air pressure and temperature are typical high-dimensional data and have a lot of redundancy. At present, the scenario method and interval method generally cluster the output of new energy directly, which fails to reflect the potential impact of meteorological factors and is difficult to accurately characterize the output of new energy. In order to reflect the impact of meteorological factors and reduce the redundancy caused by other data, we first use KPCA to reduce the dimension of {wind power output, meteorology} data.

[0012] The KPCA dimensionality reduction method is as follows: First, the data in the sample is standardized to obtain a new sample set after standardization as shown in formula (1):

[0013] x i (i=1,2,···,k,x i ∈R k ) (1)

[0014] In the formula, x i are samples of different categories, k is the total number of samples; R k is a k-dimensional vector space;

[0015] Through nonlinear mapping Map the original data matrix X to the high-dimensional feature space and transform the problem into a solution as shown in formula (2):

[0016] Cv=λv (2)

[0017] In the formula, C is the covariance matrix in the feature space; λ is the eigenvalue; v is the eigenvector corresponding to the covariance matrix;

[0018] The vector v can be represented by the linear representation of all samples, as shown in formula (3):

[0019]

[0020] In the formula, n is the dimension of the original data; is the sample vector x i Mapping in high-dimensional space; α i for The corresponding coefficients;

[0021] Substituting formula (3) into formula (2) and introducing the kernel matrix, the simplified formula is shown in formula (4):

[0022] Kα=nλα (4)

[0023] Where K is the kernel matrix; α is the eigenvector of the kernel matrix;

[0024] Arrange λ according to size and calculate the contribution, as shown in formula (5):

[0025]

[0026] In the formula, r is the contribution rate, m is the number of principal components to be extracted, and the number of principal components with cumulative contribution above 85% is selected; i is the i-th eigenvalue. The dimensionality reduction method of photovoltaic data based on KPCA is the same as above, where photovoltaic data includes historical photovoltaic output, solar radiation and temperature.

[0027] Furthermore, step 2 is specifically as follows: using agglomerative hierarchical clustering AHC technology to classify the data after dimensionality reduction of new energy, classify the corresponding original output of new energy into a category according to the data classification, and cluster typical scenarios; the AHC hierarchical clustering algorithm regards various types of irregular data as an independent cluster, merges two similar clusters, and terminates the merger when the required clustering results are met. Agglomerative hierarchical clustering is a bottom-up clustering method, which has a good clustering effect on wind power output data under different complex working conditions. Compared with divisive hierarchical clustering and other prototype clustering algorithms, the AHC algorithm has high clustering accuracy, high anti-interference ability for abnormal wind power data and low noise sensitivity, and high algorithm efficiency. The specific steps are as follows:

[0028] Step 2.1: Set the sample set X = (x1, x2, x3, ..., x k ) is considered as a single cluster;

[0029] Step 2.2: Calculate the average linkage distance between two clusters to form a distance matrix, where the average linkage distance calculation is shown in formula (6):

[0030]

[0031] Where, d avg (c i ,c j ) is cluster c i With cluster c j The average linkage distance between ||x-x'|| is the Euclidean distance between feature vectors x and x'; k i and k j For cluster c i and c j The number of eigenvectors in ;

[0032] Step 2.3: In each iteration, two clusters with the smallest distance are merged into one cluster;

[0033] Step 2.4, repeat step 2.3 until all data points are merged into one cluster, predetermine the number of clusters, and finally determine the typical scenarios of wind power and photovoltaic power based on the results obtained by the AHC hierarchical clustering algorithm;

[0034] Step 2.5: After using the KPCA method to obtain the dimensionally reduced wind power and photovoltaic data, the AHC method is used to cluster the new energy output characteristics into four clusters of spring, summer, autumn and winter seasons to obtain four groups of typical output scenarios. The four groups of new energy output data are used as new energy output prediction values ​​for EHIES optimization scheduling.

[0035] Furthermore, step 3 specifically includes the following steps:

[0036] Step 3.1, Goal 1: Minimize the total system operating cost;

[0037] The operating costs mainly include the operating costs of batteries, combined power generation units, wind power generation, heat storage tanks and gas boilers; the operating costs of combined power generation units and gas boilers, and the electricity interaction costs with the main grid;

[0038] M obj1 =M OP +M CHP +M GB +M Grid (7)

[0039] Where M obj1 is the total operating cost, M OP is the operating cost of the power grid, M CHP is the operating cost of the combined power generation unit, M GBis the operating cost of the gas boiler, M Grid is the grid interaction cost, yuan; among them,

[0040]

[0041] In the formula, is the output of the new energy at time t, n RES Indicates the number of new energy stations; is the output of the combined power generation unit, n CHP Indicates the number of CPG units; and are the battery charging and discharging power, n ES Indicates the number of batteries; and are the heat storage tank charge and discharge power, n TST Indicates the number of heat storage tanks; is the heat release power of the gas boiler, n GB Indicates the number of gas boilers; α RES is the operation and maintenance cost coefficient of the new energy unit, α CHP is the operation and maintenance cost coefficient of the combined power generation unit, α ES is the battery operation and maintenance cost coefficient, α TST is the operating cost coefficient of the heat pump, α GB is the operating cost coefficient of the gas unit, yuan;

[0042]

[0043] In the formula, and are the power purchased from the power grid and the power sold to the power grid respectively; c Gas is the price of purchased natural gas for the combined power generation unit and gas boiler, which is 0.349 yuan / kW·h. and is the operating efficiency of the combined power generation unit and the gas unit, c Grid is the electricity purchase price from the power grid, yuan / kW·h;

[0044] Step 3.2, Goal 2: Minimize the total carbon emissions of the system;

[0045] System carbon emissions mainly include CO2 generated by combined power generation units, power purchases from the grid, and gas boilers;

[0046]

[0047] In the formula, C obj2 is the total carbon emissions of the system, C CHP is the carbon emissions generated by the combined power generation unit, C Grid Carbon emissions from purchasing electricity from the grid, C GBis the carbon emission generated by the gas boiler, kg; is the carbon emission coefficient of the combined power generation unit, is the carbon emission coefficient of electricity purchased from the power grid, is the carbon emission coefficient of the gas unit, kg / kW·h;

[0048] The constraints of the regional power grid are as follows:

[0049]

[0050] In the formula, It is the operation status of the combined power generation unit, 0-off, 1-on; I bus is the busbar node set; and are the upper and lower limits of the combined power generation unit output, respectively, and T=24 is the dispatching time;

[0051]

[0052] In the formula, is the waste heat boiler efficiency, is the heat loss rate, is the heat generation power of the combined power generation unit, kW;

[0053]

[0054] In the formula, As the upper limit of its power generation capacity, is the upper limit of GB heating power, kW;

[0055]

[0056] In the formula, The power consumption of electric boiler, As its power limit, For the operating efficiency of electric boiler, is the heating power of the electric boiler, kW;

[0057]

[0058] In the formula, and They are the battery charging and discharging states respectively; and are the charging and discharging powers of the battery at time t, is the self-discharge efficiency of the battery, and are the charging and discharging efficiencies of the battery, is the electric energy stored in the battery at time t, kW;

[0059]

[0060]

[0061] In the formula, and are the active and reactive powers of the branch respectively, is the branch power limit, is the system electrical load, is the system heat load, kW; and are branch resistance and reactance respectively, V i,t is the bus voltage, V0 is the reference voltage, I BR For branch collection;

[0062] Step 3.3: Use the normalized plane constraint method NNC to solve the multi-objective optimization problem, obtain the uniformly distributed Pareto frontier, and write the multi-objective optimization model of system operation cost M(X) and carbon emission C(X) into the following compact form:

[0063]

[0064] In the formula, g(X) is the equality constraint in the original model, h(X) is the inequality constraint in the original model, and the NNC method adds constraints to the original multi-objective optimization problem, converting it into a series of single-objective optimization problems. Each time a single-objective optimization problem is solved, a point can be obtained, thereby forming the Pareto frontier. The optimization objectives of the present invention are 2, and the NNC method can be intuitively explained in a rectangular coordinate system.

[0065] Furthermore, in step 3.3, the normalized normal plane constraint method NNC solves the multi-objective optimization problem, specifically comprising the following steps:

[0066] Step 3.3.1. Solve M separately obj1 and C obj2 , we get the two endpoints of the Pareto frontier, and when we solve the minimum cost separately, we get M min and C max , solve the carbon emission minimum separately to get C min and M max ;

[0067] Step 3.3.2: Normalize the two objective functions to obtain the cost and carbon emission coordinate axes. Points (0,1) and (1,0) are the points where M and C take the minimum value; the straight line between the two points is called the Utopia line;

[0068] Step 3.3.3, select any point f on the Utopia line pAnd draw a vertical line, along which point A on the Pareto frontier corresponds. When the number of constraints increases, the objective function can only take values ​​in the area above this vertical line, and then C is calculated separately. obj2 To find the minimum, we can get point A. After adding constraints, the original model can be transformed into:

[0069]

[0070] In the formula, the last inequality in the constraint condition means that the objective function vector is constrained to be Figure 3 In the feasible domain shown in the figure, w is a series of uniform values ​​in the interval [0,1], which determines the point f p At the position on the Utopia line, let the number of Pareto optimal points calculated be m, where m includes two endpoints, then the value of w is 0, 1 / (m-1), 2 / (m-1), …, (m-2) / (m-1), 1;

[0071] Step 3.3.4: Repeat step 3.3.3 to obtain a series of uniformly distributed Pareto optimal points, thereby depicting the Pareto frontier and obtaining the Pareto optimal solution set for the multi-objective optimization problem.

[0072] Furthermore, step 4 is specifically as follows:

[0073] Emission reduction cost is defined as the potential cost of reducing unit CO2. Excessive emission reduction cost will lead to passive emission reduction in the power industry and hinder the low-carbon development of the power system. Quantitative evaluation of emission reduction cost can more accurately express the relationship between emission reduction cost and emission reduction level. If the carbon emission reduction of the system increases, the corresponding total operating cost will also increase. In this case, the relationship between unit emission reduction and operating cost is the emission reduction cost, which can be expressed as:

[0074]

[0075] In the formula, is the carbon reduction cost of the integrated energy system; M obj1,P+1 , M obj1,P are two adjacent points on the Pareto frontier of system operation cost; C obj2,P+1 , C obj2,P are two adjacent points on the Pareto frontier of system carbon emissions.

[0076] The present invention also discloses an emission reduction cost evaluation system for an electric-thermal integrated energy system based on KPCA-AHC new energy dimensionality reduction clustering, comprising:

[0077] Design unit: Use the kernel principal component analysis method to reduce the dimension of new energy data, and cluster the reduced data of new energy through the agglomerative hierarchical clustering method to obtain the typical output scenarios of wind power and photovoltaic power; take the Electricity-Heat Integrated Energy Systems (EHIES) as the research object, and establish a multi-objective optimization scheduling model with the minimum operating cost and carbon emissions; use the normalized plane constraint method to solve the model, obtain the system operating cost and carbon emission Pareto frontier, and evaluate the system's emission reduction cost based on the system cost-carbon emission Pareto frontier;

[0078] Modeling unit: Establish a new energy dimension reduction clustering model based on the KPCA-AHC method, including the KPCA dimension reduction model of new energy data and the AHC clustering model of new energy output; establish a multi-objective optimization scheduling model for the electric and thermal integrated energy system, with objective function 1 minimizing the system's operating cost and objective function 2 minimizing the system's carbon emissions. Build regional power grid and heat network models respectively to achieve an accurate description of the power system flow and heat network heat flow; use the NNC method to solve the multi-objective optimization problem and obtain the system's carbon emissions and operating cost Pareto frontiers;

[0079] Calculation unit: After obtaining the Pareto optimal solution set of the multi-objective optimization problem, the slopes of different optimal points are calculated using formula (25) according to the Pareto frontier of system operation cost and CO2 emission, so as to measure the carbon emission reduction cost of the system.

[0080] The present invention also discloses a computer device, comprising a memory, a processor and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method of the present invention.

[0081] The present invention also discloses a computer-readable storage medium on which a computer program / instruction is stored. When the computer program / instruction is executed by a processor, the steps of the method of the present invention are implemented.

[0082] The present invention also discloses a computer program product, comprising a computer program / instruction, which implements the steps of the method of the present invention when executed by a processor.

[0083] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages:

[0084] The present invention uses the KPCA method to reduce the dimension of multi-dimensional historical data of wind power and photovoltaic power, which can reduce the redundancy of historical data and better characterize the output characteristics of new energy. Using the AHC method, the reduced-dimensional data is clustered based on the distance and similarity of the clusters, which can optimize the classification of new energy and more accurately characterize the uncertainty of new energy output. The new energy output data obtained by dimensionality reduction clustering is used as the new energy prediction value, and the multi-objective optimization scheduling of EHIES is solved by the NNC method to obtain the Pareto frontier of system cost and carbon emission. It can be seen that the operating cost of EHIES and the carbon emission change trend are opposite; the emission reduction cost of EHIES is calculated based on the Pareto frontier of system operating cost and carbon emission. The example shows that the emission reduction cost varies in the range of [0.184 yuan, 0.895 yuan], which provides a reference and reference for the low-carbon scheduling and carbon emission reduction evaluation of EHIES. BRIEF DESCRIPTION OF THE DRAWINGS

[0085] Figure 1 This is a flow chart of an emission reduction cost assessment method for an electric-thermal integrated energy system based on KPCA-AHC new energy dimensionality reduction clustering according to an embodiment of the present invention.

[0086] Figure 2 It is an electric and thermal integrated energy system structure.

[0087] Figure 3 It is the electricity price data of the electric-thermal integrated energy system.

[0088] Figure 4 It is the outdoor temperature data of the area where the electric and thermal integrated energy system is located.

[0089] Figure 5 This is the result of wind power-photovoltaic dimensionality reduction. Figure 5 (a) is the original wind power data and its distribution after dimension reduction; Figure 5 (b) is the original photovoltaic data and its data distribution after dimensionality reduction; Figure 5 The left figure of (c) shows the cumulative contribution of each main element of wind power, and the right figure shows the cumulative contribution of each main element of photovoltaic power.

[0090] Figure 6 This is the result of wind power-photovoltaic dimensionality reduction clustering. Figure 6 (a) and Figure 6 (b) shows the typical wind power and photovoltaic output scenarios obtained by AHC clustering after KPCA dimensionality reduction, where the black line is the historical output and the red line is the typical output. Figure 6 (c) shows four typical scenarios of new energy output.

[0091] Figure 7 This is the schematic diagram of the NNC method.

[0092] Figure 8 The Pareto result of cost-carbon emission.

[0093] Fig. 9 Optimize scheduling results for EHIES.

[0094] Fig.10 The result is the cost of emission reduction. DETAILED DESCRIPTION

[0095] The following is combined with Figure 1 The present invention is further described in detail with reference to the process of FIG.

[0096] Example 1

[0097] Attach Figure 2 Taking the electric-heat integrated energy system shown in the figure as an example, the system consists of an IEEE 33-node distribution network and a 51-node regional heating network. The heating network mainly relies on the waste heat of the combined heat generation unit and the gas boiler for heating; the black line area in the figure represents the distribution network, and the red line area represents the heating network. ES represents batteries, RES represents new energy, CHP represents the combined heat and electricity generation unit, GB represents the gas boiler, and HX represents the heat exchanger; "P1 / 45 / 0.6" represents the heating network pipeline 1, with a length of 45m and a pipe diameter of 0.6m; the scheduling frequency is set to 1h / 1 time, and the scheduling time is 24 hours. The equipment capacity and parameters in the system are shown in Table 1, of which the data of the IEEE-33-node distribution network and the 51-node heating network are shown in Table 2, and the electricity price data of the electric-heat integrated energy system are shown in the attached Figure 3 , the outdoor temperature data of the area where the electric heating integrated energy system is located is attached Figure 4 In order to meet the energy comfort of users, the heating network ensures that the indoor temperature of users is controlled at 18-23°C. The simulation platform is Matlab2018b+Yalmip+Cplex12.8, and the computer configuration is Corei7-55600CPU@3.50GHz 16GB RAM.

[0098] Table 1 Equipment capacity and parameters

[0099]

[0100]

[0101] Table 2 Heating network pipeline parameters

[0102]

[0103]

[0104] S1, using the kernel principal component analysis method to reduce the dimension of the new energy data, based on the reduced dimension data, clustering the new energy output through the agglomerative hierarchical clustering method, and obtain the typical output scenarios of wind power and photovoltaic power;

[0105] The proposed KPCA-AHC method for generating a set of typical output scenarios for new energy sources mainly includes three core steps:

[0106] 1) Dimensionality reduction of new energy data based on KPCA;

[0107] The following is an explanation using wind power as an example. Wind power output is not only strongly correlated with historical output, but also closely related to meteorological factors such as wind speed, wind direction, air pressure and temperature. The historical data composed of wind power output, wind speed, wind direction, air pressure and temperature are typical high-dimensional data and contain a lot of redundancy. At present, the scenario method and interval method generally cluster the output of new energy directly, which fails to reflect the potential impact of meteorological factors and makes it difficult to accurately characterize the output of new energy. In order to reflect the impact of meteorological factors and reduce the redundancy caused by other data, KPCA is first used to reduce the dimension of wind power and meteorological data. The KPCA dimensionality reduction method is as follows:

[0108] First, the data in the sample is standardized to obtain a new sample set after standardization as shown in formula (1):

[0109] x i (i=1,2,···,k,x i ∈R k ) (1)

[0110] Where: x i are samples of different categories, k is the total number of samples; R k is a k-dimensional vector space.

[0111] Through nonlinear mapping Map the original data matrix X to the high-dimensional feature space and transform the problem into a solution as shown in formula (2):

[0112] Cv=λv (2)

[0113] Where: C is the covariance matrix in the feature space; λ is the eigenvalue; v is the eigenvector corresponding to the covariance matrix.

[0114] The vector v can be represented by the linear representation of all samples, as shown in formula (3):

[0115]

[0116] Where: n is the dimension of the original data; is the sample vector x i Mapping in high-dimensional space; α i for The corresponding coefficient.

[0117] Substituting formula (3) into formula (2) and introducing the kernel matrix, the simplified formula is shown in formula (4):

[0118]

[0119] Where: K is the kernel matrix; α is the eigenvector of the kernel matrix.

[0120] Arrange λ according to size and calculate the contribution, as shown in formula (5):

[0121]

[0122] In the formula: r is the contribution rate, m is the number of principal components to be extracted, and generally the number of principal components with a cumulative contribution of more than 85% is selected; λ i is the ith eigenvalue.

[0123] The dimensionality reduction method of photovoltaic data based on KPCA is the same as above and will not be repeated here. The photovoltaic data includes historical photovoltaic output, solar radiation and temperature. Based on one year of wind power historical data (including wind power output, wind speed, wind direction, air pressure, temperature) and photovoltaic historical data (including photovoltaic output, solar radiation and temperature), the KPCA method is used to reduce the dimension of wind power and photovoltaic historical data. The results are shown in the attached Figure 5 .in, Figure 5 (a) and Figure 5 The left figure in (b) shows the original data distribution of wind power and photovoltaic power, and the right figure shows the data distribution after dimensionality reduction; Figure 5 The left figure of (c) shows the contribution of each main element of wind power, and the right figure shows the contribution of each main element of photovoltaic power.

[0124] From the attached Figure 5 It can be seen that:

[0125] (1) Before dimensionality reduction, the three features are high-dimensional and nonlinear, and the features affect each other, making it impossible to decouple the analysis. After KPCA dimensionality reduction, the three principal components have a significant linear relationship and can be decoupled.

[0126] (2) From the attachment Figure 5 (c) It can be seen that when the first three principal components are selected for wind power, the cumulative contribution is 92.02%, and when the first four principal components are selected, the contribution increases to 95.84. The increase in one-dimensional principal component but the cumulative contribution increases slightly. In order to ensure that the dimensionality reduction data is less distorted and meet the needs of data processing scale, the number of dimensionality reduction principal components of wind power is selected to be 3. Similarly, photovoltaic power has a total of 3-dimensional data (output, temperature, radiation). When the number of dimensionality reduction principal components of photovoltaic power is selected to be 2, its cumulative contribution rate is 94.94%. The number of dimensionality reduction principal components of photovoltaic power can characterize the original characteristics of photovoltaic data.

[0127] 2) Using AHC hierarchical clustering technology, the data after dimension reduction of new energy is classified, the corresponding original output of new energy is classified into one category according to the data classification, and typical scenarios are clustered;

[0128] The AHC hierarchical clustering algorithm treats various types of irregular data as an independent cluster, merges two similar clusters, and terminates the merger when the required clustering results are met. Agglomerative hierarchical clustering is a bottom-up clustering method that has a good clustering effect on wind power output data under different complex working conditions. Compared with divisive hierarchical clustering and other prototype clustering algorithms, the AHC algorithm has high clustering accuracy, high anti-interference ability for abnormal wind power data, low noise sensitivity, and high algorithm efficiency. Therefore, this model uses the AHC algorithm to cluster wind power output scenarios. The specific process is as follows:

[0129] (1) Let the sample set X = (x1, x2, x3, ..., x k ) is considered as a single cluster.

[0130] (2) Calculate the average linkage distance between two clusters to form a distance matrix. The average linkage distance calculation is shown in formula (6):

[0131]

[0132] Where: d avg (c i ,c j ) is cluster c i With cluster c j The average linkage distance between ||x-x'|| is the Euclidean distance between feature vectors x and x'; k i and k j For cluster c i and c j The number of eigenvectors in .

[0133] (3) In each iteration, two clusters with the smallest distance are merged into one cluster.

[0134] (4) Repeat step (3) until all data points are merged into one cluster. The number of clusters is determined in advance, and finally the typical scenarios of wind power and photovoltaic power are determined based on the results obtained by the AHC hierarchical clustering algorithm.

[0135] 3) According to the seasonal output characteristics of new energy sources, a set of typical wind and solar output scenarios is obtained.

[0136] After using the KPCA method to obtain the dimension-reduced wind power and photovoltaic data, the AHC method is used to cluster the output characteristics according to the spring, summer, autumn and winter seasons (4 clusters) of new energy, and 4 groups of typical output scenarios are obtained. See the attached Figure 6 .

[0137] Four sets of new energy (wind power and photovoltaic) output data are used as new energy output forecast values ​​for EHIES optimization scheduling.

[0138] S2, with the goal of minimizing operating costs and carbon emissions, establishes a multi-objective optimization scheduling model for the electric-thermal integrated energy system, uses the normalized plane constraint method to solve the model, and obtains the Pareto frontier of system operating costs and carbon emissions;

[0139] Objective 1: Minimize the total system operating cost.

[0140] The operating costs mainly include the operating costs of batteries, combined power generation units, wind power generation, heat storage tanks and gas boilers; the operating costs of combined power generation units and gas boilers, and the electricity interaction costs with the main grid.

[0141] M obj1 =M OP +M CHP +M GB +M Grid (7)

[0142] Where M obj1 is the total operating cost, M OP is the operating cost of the power grid, M CHP is the operating cost of the combined power generation unit, M GB is the operating cost of the gas boiler, M Grid is the grid interaction cost, RMB.

[0143]

[0144] In the formula, is the output of the new energy at time t, n RES Indicates the number of new energy stations; is the output of the combined power generation unit, n CHP Indicates the number of CPG units; and are the battery charging and discharging power, n ES Indicates the number of batteries; and are the heat storage tank charge and discharge power, n TST Indicates the number of heat storage tanks; is the heat release power of the gas boiler, n GB Indicates the number of gas boilers; α RES is the operation and maintenance cost coefficient of the new energy unit, α CHP is the operation and maintenance cost coefficient of the combined power generation unit, α ES is the battery operation and maintenance cost coefficient, α TST is the operating cost coefficient of the heat pump, α GB is the operating cost coefficient of the gas unit, yuan;

[0145]

[0146] In the formula, and are the power purchased from the power grid and the power sold to the power grid respectively; c Gas is the price of purchased natural gas for the combined power generation unit and gas boiler, which is 0.349 yuan / kW·h. and is the operating efficiency of the combined power generation unit and the gas unit, c Grid is the electricity purchase price from the power grid, yuan / kW·h;

[0147] Goal 2: Minimize the total carbon emissions of the system.

[0148] The system’s carbon emissions mainly include CO2 generated by combined power generation units, power purchases from the grid and gas boilers.

[0149]

[0150] In the formula, C obj2 is the total carbon emissions of the system, C CHP is the carbon emissions generated by the combined power generation unit, C Grid Carbon emissions from purchasing electricity from the grid, C GB is the carbon emission generated by the gas boiler, kg; is the carbon emission coefficient of the combined power generation unit, is the carbon emission coefficient of electricity purchased from the power grid, is the carbon emission coefficient of the gas unit, kg / kW·h.

[0151] The constraints of the regional power grid are as follows:

[0152] In the formula, It is the operation status of the combined power generation unit, 0-off, 1-on; I bus is the busbar node set; and are the upper and lower limits of the combined power generation unit output, respectively, and T=24 is the dispatching time;

[0153]

[0154] In the formula, is the waste heat boiler efficiency, is the heat loss rate, is the heat generation power of the combined power generation unit, kW;

[0155]

[0156] In the formula, As the upper limit of its power generation capacity, is the upper limit of GB heating power, kW;

[0157]

[0158] In the formula, The power consumption of electric boiler, As its power limit, For the operating efficiency of electric boiler, is the heating power of the electric boiler, kW;

[0159]

[0160] In the formula, and They are the battery charging and discharging states respectively; and are the charging and discharging powers of the battery at time t, is the self-discharge efficiency of the battery, and are the charging and discharging efficiencies of the battery, is the electric energy stored in the battery at time t, kW;

[0161]

[0162] In the formula, and are the active and reactive powers of the branch respectively, is the branch power limit, is the system electrical load, is the system heat load, kW; and are branch resistance and reactance respectively, V i,t is the bus voltage, V0 is the reference voltage, I BR For branch collection;

[0163] Normalized Normal Constraint (NNC) is an effective method for solving multi-objective optimization problems and can obtain a uniformly distributed Pareto frontier. The multi-objective optimization model of system operating cost M(X) and carbon emissions C(X) is written in the following compact form:

[0164]

[0165] In the formula, g(X) is the equality constraint in the original model, and h(X) is the inequality constraint in the original model. The NNC method adds constraints to the original multi-objective optimization problem and transforms it into a series of single-objective optimization problems. Each time a single-objective optimization problem is solved, a point can be obtained, thus forming the Pareto frontier. The optimization objectives of this model are two, so the NNC method can be intuitively explained in the rectangular coordinate system. The principle of the NNC method is shown in the attached figure. Figure 7 shown.

[0166] NNC method steps:

[0167] 1) Solve M separately obj1 and C obj2 , we get the two endpoints of the Pareto frontier, and when we solve the minimum cost separately, we get M min and C max , solve the carbon emission minimum separately to get C min and M max ;

[0168] 2) Normalize the two objective functions to obtain the cost and carbon emission coordinate axes. Points (0,1) and (1,0) are the points where M and C take the minimum value; the straight line between the two points is called the Utopia line;

[0169] 3) Choose any point f on the Utopia line p And draw a vertical line, along which point A on the Pareto frontier corresponds. When the number of constraints increases, the objective function can only take values ​​in the area above this vertical line, and then C is calculated separately. obj2 To find the minimum, we can get point A. After adding constraints, the original model can be transformed into:

[0170]

[0171] In the formula, the last inequality in the constraint condition means that the objective function vector is constrained to be Figure 7 In the feasible domain shown in the figure, w is a series of uniform values ​​in the interval [0,1], which determines the point f p At the position on the Utopia line, assuming that the number of calculated Pareto optimal points is m (including the two endpoints), the value of w is 0, 1 / (m-1), 2 / (m-1), …, (m-2) / (m-1), 1.

[0172] 4) Repeat step 3) to obtain a series of uniformly distributed Pareto optimal points, thereby depicting the Pareto frontier. After obtaining the Pareto optimal solution set of the multi-objective optimization problem, the emission reduction cost model is used to calculate the slopes of different optimal points, and the carbon emission reduction cost of the system is calculated accordingly.

[0173] Based on the obtained four sets of new energy typical output scenario data, EHIES multi-objective optimization scheduling is carried out (new energy scenarios 1-4 correspond to the example scenarios 1-4 here), and the NNC method is used to solve the multi-objective optimization scheduling model, and four sets of cost-carbon emission optimization results are obtained, as shown in Table 3. Figure 8The EHIES multi-objective optimization cost-carbon emission Pareto frontier solved by the NNC method shows that the system carbon emissions and operating costs have opposite trends. This is because in order to reduce its own carbon emissions, the system needs to optimize its operating strategy or take economic measures, which leads to an increase in its own operating costs.

[0174] Table 3 EHIES cost-carbon emission results

[0175]

[0176] As can be seen from Table 3, the power purchase of the power grid in the four scenarios is 0, indicating that when the power source can meet its own load demand, power generation is given priority to meet the power balance of the park.

[0177] From the perspective of economic cost, since the power of new energy in new energy scenarios 1-4 decreases successively, its consumption volume decreases accordingly. The system relies on combined heat generation units to meet electricity and heat demand, and the operating costs of the system and combined heat generation units increase accordingly.

[0178] From the perspective of carbon emissions, as the penetration rate of new energy decreases, the system's carbon emissions gradually increase, because the decrease in new energy power generation leads to an increase in power generation of the combined power generation unit, and carbon emissions increase accordingly; the combined power generation unit can supply electricity and heat to the system at the same time, and the heat generated by its power generation is used for system heating, reducing the output of the gas boiler, and its carbon emissions decrease accordingly.

[0179] Attached Fig. 9 The figure shows the results of the EHIES electric heat balance optimization scheduling. It can be seen that due to the carbon cost constraints of the system, the system mainly relies on combined heat generation units and new energy power generation to ensure the power balance of the system, and the thermal balance mainly relies on the waste heat of the combined heat generation units and the heat generated by the gas boiler. In the four groups of scenarios, the penetration rate of new energy gradually decreases, and the battery charge and discharge gradually increase and are more concentrated. The battery discharge is mainly concentrated in the period of 18:00-21:00, which is the peak period of power consumption of the system. The battery charging of scenarios 1-3 is mainly concentrated in the period of 11:00-16:00, and the battery charging of scenario 4 is mainly concentrated in the period of 00:00-06:00, which is the low electricity consumption period of the system; when the output of the combined heat generation unit is large, the output of the gas boiler is reduced accordingly, because the waste heat of the combined heat generation unit partially meets the heat demand of the user. The combined heat generation unit can simultaneously ensure the power supply and heat supply of the system and improve the gas utilization efficiency of the system; due to the consideration of the energy comfort of the user, the external temperature and the economy of the system operation, the heat demand of the system fluctuates; the heat storage tank has a short heat release period, and the system only releases heat at 00:00 and 23:00-24:00.

[0180] S3, evaluates the system’s emission reduction costs based on the system cost-carbon emission Pareto frontier;

[0181] Emission reduction cost is defined as the potential cost of reducing unit CO2. Excessive emission reduction cost will lead to passive emission reduction in the power industry, hindering the low-carbon development of the power system. Quantitative evaluation of emission reduction cost can more accurately express the relationship between emission reduction cost and emission reduction level. If the carbon emission reduction of the system increases, the corresponding total operating cost will also increase. In this case, the relationship between unit emission reduction and operating cost is the emission reduction cost, which can be expressed by the formula:

[0182]

[0183] In the formula, is the carbon reduction cost of the integrated energy system; M obj1,P+1 , M obj1,P are two adjacent points on the Pareto frontier of system operation cost; C obj2,P+1 , C obj2,P are two adjacent points on the Pareto frontier of system carbon emissions.

[0184] Attached Fig.10 The emission reduction cost results of the electric-thermal integrated energy system under four scenarios show that the emission reduction costs of scenarios 1 and 2 are on the rise, while the system emission reduction costs of scenarios 3 and 4 are on the decline, which is closely related to the change rate of the system's cost-carbon emission Pareto frontier; the maximum emission reduction cost of the system in the example is 0.895 yuan / kgCO2, and the minimum emission reduction cost is 0.184 yuan / kgCO2, indicating that the system operating cost varies within this range. This method can be used as a reference for studying the emission reduction cost of integrated energy systems.

[0185] In a second aspect, the present invention provides an emission reduction cost evaluation system for an electric-thermal integrated energy system based on KPCA-AHC new energy dimensionality reduction clustering, comprising:

[0186] Design unit: Use the kernel principal component analysis method to reduce the dimension of new energy data, and cluster the reduced data of new energy through the agglomerative hierarchical clustering method to obtain the typical output scenarios of wind power and photovoltaics; take the electricity-heat integrated energy system (EHIES) as the research object, and establish a multi-objective optimization scheduling model with minimum operating cost and carbon emissions; use the normalized plane constraint method to solve the model, obtain the system operating cost and carbon emission Pareto frontier, and evaluate the system's emission reduction cost based on the system cost-carbon emission Pareto frontier.

[0187] Modeling unit: Establish a new energy dimension reduction clustering model based on the KPCA-AHC method, including the KPCA dimension reduction model of new energy data and the AHC clustering model of new energy output; establish a multi-objective optimization scheduling model for the electric and thermal integrated energy system, with objective function 1 minimizing the system's operating cost and objective function 2 minimizing the system's carbon emissions. Build regional power grid and heat network models respectively to achieve an accurate description of the power system flow and heat network heat flow; use the NNC method to solve the multi-objective optimization problem and obtain the system's carbon emissions and operating cost Pareto frontiers;

[0188] Calculation unit: After obtaining the Pareto optimal solution set of the multi-objective optimization problem, the slopes of different optimal points are calculated using formula (25) according to the Pareto frontier of system operation cost and CO2 emission, so as to measure the carbon emission reduction cost of the system.

Claims

1. A method for evaluating emission reduction costs of electric thermal energy systems based on KPCA-AHC new energy dimensionality reduction clustering, characterized in that: The steps include: Step 1: Collect new energy data, and use the kernel principal component analysis method KPCA to reduce the dimension of the new energy data to obtain the reduced-dimensional data; Step 2: Based on the data after dimensionality reduction, the output of renewable energy is clustered by using the agglomerative hierarchical clustering method AHC to obtain the typical output scenarios of wind power and photovoltaic power; Step 3: With the goal of minimizing operating costs and carbon emissions, a multi-objective optimization scheduling model for the electric-thermal integrated energy system is established, and the normalized plane constraint method is used to solve the model to obtain the Pareto frontier of system operating costs and carbon emissions; Step 4: Based on the obtained Pareto frontier, evaluate the carbon reduction cost of the electric thermal energy system.

2. According to claim 1, a method for evaluating emission reduction costs of electric thermal energy systems based on KPCA-AHC new energy dimensionality reduction clustering is characterized in that: Step 1 is as follows: The new energy data includes wind power data and photovoltaic data; The KPCA dimensionality reduction method is as follows: First, the data in the sample is standardized to obtain a new sample set after standardization as shown in formula (1): x i (i=1,2,···,k,x i ∈R k ) (1) In the formula, x i are samples of different categories, k is the total number of samples; R k is a k-dimensional vector space; Through nonlinear mapping The original data matrix X = [x1, x2, x3, ..., x k ] is mapped to a high-dimensional feature space, and the problem is transformed into solving the equation shown in formula (2): Cv=λv (2) In the formula, C is the covariance matrix in the feature space; λ is the eigenvalue; v is the eigenvector corresponding to the covariance matrix; The vector v can be represented by the linear representation of all samples, as shown in formula (3): In the formula, n is the dimension of the original data; is the sample vector x i Mapping in high-dimensional space; α i for The corresponding coefficients; Substituting formula (3) into formula (2) and introducing the kernel matrix, the simplified formula is shown in formula (4): Kα=nλα (4) Where K is the kernel matrix; α is the eigenvector of the kernel matrix; Arrange λ according to size and calculate the contribution, as shown in formula (5): In the formula, r is the contribution rate, m is the number of principal components to be extracted, and the principal components with cumulative contribution above 85% are selected; i is the ith eigenvalue.

3. According to claim 1, a method for evaluating emission reduction costs of electric thermal energy systems based on KPCA-AHC new energy dimensionality reduction clustering is characterized in that: Step 2 is as follows: using the agglomerative hierarchical clustering (AHC) technology to classify the data after dimension reduction of new energy, classify the corresponding original output of new energy according to the data classification results, and cluster typical scenarios; the specific steps are as follows: Step 2.1: Set the sample set X = [x1, x2, x3, ..., x k Each type of sample data in ] is regarded as a single cluster; Step 2.2: Calculate the average linkage distance between any two clusters to form a distance matrix, where the average linkage distance calculation is shown in formula (6): Where, d avg (c i ,c j ) is cluster c i With cluster c j The average linkage distance between ||x-x'|| is the Euclidean distance between feature vectors x and x'; k i and k j For cluster c i and c j The number of samples in Step 2.3: In each iteration, two clusters with the smallest distance are merged into one cluster; Step 2.4, repeat step 2.3 until all samples are merged into one cluster, predetermine the number of clusters, and finally determine the typical scenarios of wind power and photovoltaic power based on the results obtained by the AHC hierarchical clustering algorithm; Step 2.5: After using the KPCA method to obtain the dimensionally reduced wind power and photovoltaic data, the AHC method is used to cluster the new energy output characteristics into four clusters of spring, summer, autumn and winter seasons to obtain four groups of typical output scenarios. The four groups of new energy output data are used as new energy output prediction values ​​for EHIES optimization scheduling.

4. According to claim 1, a method for evaluating emission reduction costs of electric thermal energy systems based on KPCA-AHC new energy dimensionality reduction clustering is characterized in that: Step 3 specifically includes the following steps: Step 3.1, Goal 1: Minimize the total system operating cost; The operating costs mainly include the operating costs of batteries, combined power generation units, wind power generation, heat storage tanks and gas boilers; the operating costs of combined power generation units and gas boilers, and the electricity interaction costs with the main grid; M obj1 =M OP +M CHP +M GB +M Grid (7) Where M obj1 is the total operating cost of the integrated energy system, M OP is the operating cost of the power grid, M CHP is the operating cost of the combined power generation unit, M GB is the operating cost of the gas boiler, M Grid is the grid interaction cost, yuan; among them, In the formula, is the output of the new energy at time t, n RES Indicates the number of new energy stations; is the output of the combined power generation unit, n CHP Indicates the number of CPG units; and are the battery charging and discharging power, n ES Indicates the number of batteries; and are the heat storage tank charge and discharge power, n TST Indicates the number of heat storage tanks; is the heat release power of the gas boiler, n GB Indicates the number of gas boilers; α RES is the operation and maintenance cost coefficient of the new energy unit, α CHP is the operation and maintenance cost coefficient of the combined power generation unit, α ES is the battery operation and maintenance cost coefficient, α TST is the operating cost coefficient of the heat pump, α GB is the operating cost coefficient of the gas unit, yuan; In the formula, and are the power purchased and sold by the power grid respectively; c Gas is the price of purchased natural gas for the combined power generation unit and gas boiler, which is 0.349 yuan / kW·h. and is the operating efficiency of the combined power generation unit and the gas unit, c Grid is the electricity purchase price from the power grid, yuan / kW·h; Step 3.2, Goal 2: Minimize the total carbon emissions of the system; System carbon emissions mainly include CO2 generated by combined power generation units, power purchases from the grid, and gas boilers; In the formula, C CO2 is the total carbon emissions of the system, C CHP is the carbon emissions generated by the combined power generation unit, C Grid Carbon emissions from purchasing electricity from the grid, C GB is the carbon emission generated by the gas boiler, kg; is the carbon emission coefficient of the combined power generation unit, is the carbon emission coefficient of electricity purchased from the power grid, is the carbon emission coefficient of the gas unit, kg / kW·h; The constraints of the regional power grid are as follows: In the formula, and Indicates the running status of the combined power generation unit, 0-off, 1-on; I bus is the busbar node set; and are the upper and lower limits of the combined power generation unit output, respectively, and T=24 is the dispatching time; In the formula, is the waste heat boiler efficiency, is the heat loss rate, is the heat generation power of the combined power generation unit, kW; In the formula, As the upper limit of its power generation capacity, is the upper limit of GB heating power, kW; In the formula, The power consumption of electric boiler, As its power limit, For the operating efficiency of electric boiler, is the heating power of the electric boiler, kW; In the formula, and Battery charging and discharging status; is the self-discharge efficiency of the battery, and is the charging and discharging efficiency of the battery, is the electric energy stored in the battery at time t, kW; In the formula, and is the active and reactive power of the branch, and are the upper and lower limits of branch power respectively, is the system electrical load, is the system heat load, kW; and are branch resistance and reactance respectively, V i,t is the bus voltage, V0 is the reference voltage, I BR For branch collection; Step 3.3: Use the normalized plane constraint method NNC to solve the multi-objective optimization problem, obtain the uniformly distributed Pareto frontier, and write the multi-objective optimization model of the system total operating cost M(X) and carbon emissions C(X) into the following compact form: In the formula, g(X) is the equality constraint in the original model, and h(X) is the inequality constraint in the original model. The NNC method adds constraints to the original multi-objective optimization problem and transforms it into a series of single-objective optimization problems. Each time a single-objective optimization problem is solved, a point can be obtained, thus forming the Pareto frontier.

5. According to claim 4, a method for evaluating emission reduction costs of electric thermal energy systems based on KPCA-AHC new energy dimensionality reduction clustering is characterized in that: In step 3.3, the normalized normal plane constraint method NNC solves the multi-objective optimization problem, specifically including the following steps: Step 3.3.

1. Solve M separately obj1 and C obj2 , we get the two endpoints of the Pareto frontier, and when we solve the minimum cost separately, we get M min and C max , solve the carbon emission minimum separately to get C min and M max ; Step 3.3.2: Normalize the two objective functions to obtain the cost and carbon emission coordinate axes. Points (0,1) and (1,0) are the points where M and C take the minimum value; the straight line between the two points is called the Utopia line; Step 3.3.3, select any point f on the Utopia line p And draw a vertical line, along the direction of which corresponds to point A on the Pareto frontier. When the number of constraints increases, the objective function can only take values ​​in the area above this vertical line, and then C is calculated separately. CO2 To find the minimum, we can get point A. After adding constraints, the original model can be transformed into: In the formula, the last inequality in the constraint condition means that the objective function vector is constrained in the feasible domain; w is a series of uniform values ​​in the interval [0,1], which determines the point f p At the position on the Utopia line, let the number of Pareto optimal points calculated be m, where m includes two endpoints, then the value of w is 0, 1 / (m-1), 2 / (m-1), …, (m-2) / (m-1), 1; Step 3.3.4: Repeat step 3.3.3 to obtain a series of uniformly distributed Pareto optimal points, thereby depicting the Pareto frontier and obtaining the Pareto optimal solution set for the multi-objective optimization problem.

6. According to claim 1, a method for evaluating emission reduction costs of electric thermal energy systems based on KPCA-AHC new energy dimensionality reduction clustering is characterized in that: Step 4 is as follows: Emission reduction cost is defined as the potential cost of reducing unit CO2. Excessive emission reduction cost will lead to passive emission reduction in the power industry and hinder the low-carbon development of the power system. Quantitative evaluation of emission reduction cost can more accurately express the relationship between emission reduction cost and emission reduction level. If the carbon emission reduction of the system increases, the corresponding total operating cost will also increase. In this case, the relationship between unit emission reduction and operating cost is the emission reduction cost, which can be expressed as: In the formula, is the carbon reduction cost of the integrated energy system; M obj1,P+1 , M obj1,P are two adjacent points on the Pareto frontier of system operation cost; C obj2,P+1 , C obj2,P are two adjacent points on the Pareto frontier of system carbon emissions.

7. A system for evaluating emission reduction costs of electric and thermal integrated energy systems based on KPCA-AHC new energy dimensionality reduction clustering, used to implement the method according to any one of claims 1 to 6, characterized in that: include: Design unit: Use KPCA to reduce the dimension of new energy data, and cluster the reduced dimension data of new energy through AHC to obtain typical output scenarios of wind power and photovoltaic power; Taking the electric-heat interconnected integrated energy system EHIES as the research object, a multi-objective optimization scheduling model with the minimum operating cost and carbon emissions is established; the normalized plane constraint method is used to solve the model, and the system operating cost and carbon emission Pareto frontier are obtained. The emission reduction cost of the system is evaluated based on the system cost-carbon emission Pareto frontier; Modeling unit: Establish a new energy dimension reduction clustering model based on the KPCA-AHC method, including the KPCA dimension reduction model of new energy data and the AHC clustering model of new energy output; establish a multi-objective optimization scheduling model for the electric and thermal integrated energy system, with objective function 1 minimizing the system's operating cost and objective function 2 minimizing the system's carbon emissions. Build regional power grid and heat network models respectively to achieve an accurate description of the power system flow and heat network heat flow; use the NNC method to solve the multi-objective optimization problem and obtain the system's carbon emissions and operating cost Pareto frontiers; Calculation unit: After obtaining the Pareto optimal solution set of the multi-objective optimization problem, the carbon reduction cost of the system is calculated by calculating the slopes of different optimal points based on the Pareto frontier of system operating costs and CO2 emissions.

8. A computer device comprising a memory, a processor and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the steps of the method of claim 1.

9. A computer-readable storage medium having a computer program / instruction stored thereon, characterized in that: When the computer program / instructions are executed by a processor, the steps of the method according to claim 1 are implemented.

10. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the steps of the method according to claim 1 are implemented.

Images