Wind-solar carbon flow correlation analysis method and system based on PLHS sampling and sample optimization

Abstract

A wind-solar carbon flow correlation analysis method and system based on PLHS sampling and sample optimization, comprising: constructing a wind-solar output probability distribution model; constructing a system carbon flow rate comprehensive analysis model under the coupling uncertainty of wind-solar output; using a hierarchical Latin hypercube sampling method to sample wind speed and light intensity to obtain a sample sequence; an adaptive sequence search algorithm is proposed, a composite index based on spatial dispersion is used as an optimization target to optimize the sample sequence, and the sample sequence is input into the wind-solar output probability distribution model to obtain a wind-solar output sample set; traversing the sample points in the wind-solar output sample set, the system carbon flow rate and its probability distribution characteristics are calculated; for each sample point in the wind-solar output sample set, the change amount of the system node, branch and network loss carbon flow rate caused by the wind-solar output injection is calculated respectively, and the weighted average of the change amount is obtained to obtain the average influence factor of the wind-solar output on the node, branch and network loss carbon flow rate. The present application realizes reliable evaluation of system carbon emission.

Description

Technical Field

[0001] This invention belongs to the field of uncertainty analysis of carbon emission flows in power systems, and in particular relates to a method and system for wind-solar carbon flow correlation analysis based on PLHS sampling and sample optimization. Background Technology

[0002] Currently, the installed capacity of renewable energy continues to achieve new breakthroughs, maintaining its dominant position in new installations, accounting for nearly 60% of the national total. As of the first half of 2025, my country's installed renewable energy power generation capacity had exceeded 2.159 billion kilowatts, accounting for 59.2% of the country's total installed power capacity. Against this backdrop, the inherent uncertainties and intermittency of new energy units are becoming increasingly prominent, posing significant challenges to power system operation. Furthermore, the uncertainty of new energy output and the randomness of load fluctuations have a significant impact on the power flow distribution of AC and DC power systems. Since the carbon emission flow of the power system is directly related to the power flow, the uncertainty of new energy sources poses a significant challenge to the accurate quantification of the carbon emission flow of the power system.

[0003] Existing research on the impact of renewable energy integration on power system carbon flow distribution mainly focuses on carbon emission flow calculation models and allocation mechanisms, emphasizing deterministic analysis and failing to address the indirect carbon flow distribution differences caused by the uncertainty of renewable energy output. Furthermore, Monte Carlo sampling (MCS), used to characterize renewable energy uncertainty, suffers from high cost, low efficiency, and low computational accuracy. Latin Hypercube Sampling (LHS) significantly improves sampling efficiency while maintaining accuracy, more accurately characterizing the impact of wind power uncertainty on carbon flow distribution. However, existing research on LHS in characterizing renewable energy uncertainty is limited to probabilistic optimal power flow (OPF) calculations, without considering the impact of renewable energy uncertainty on carbon flow distribution.

[0004] To fully characterize the impact of new energy uncertainty on carbon flow distribution and achieve reliable assessment of system carbon emissions, this invention constructs a full-process carbon flow metering model for power systems that takes into account the stochasticity of wind and solar power output, and further proposes an incremental probabilistic carbon flow calculation method based on Progressive Latin Hypercube Sampling (PLHS). Summary of the Invention

[0005] The spatiotemporal distribution and probabilistic characteristics of wind and solar power output exhibit high uncertainty, posing a significant challenge to accurately tracking the carbon flow distribution of the system. To fully quantify the impact of new energy uncertainty on the carbon footprint and achieve a reliable assessment of system carbon emissions, it is essential to finely characterize the probabilistic distribution of this uncertainty. This invention considers the multiple stochasticities of wind and solar power and proposes a wind-solar carbon flow correlation analysis method and system based on PLHS sampling and sample optimization.

[0006] The present invention adopts the following technical solution.

[0007] This invention discloses a method for wind-solar-carbon flow correlation analysis based on PLHS sampling and sample optimization, comprising the following steps: Step 1: Construct a probability distribution model for wind and solar power output; construct a comprehensive analysis model for system carbon flow rate under the coupling uncertainty of wind and solar power output; Step 2: A hierarchical Latin hypercube sampling method is used to sample wind speed and light intensity to obtain a sample sequence; an adaptive sequence search algorithm is proposed, which uses a composite index based on spatial dispersion as the optimization objective to optimize the sample sequence. The optimized sample sequence is then input into the wind and solar power output probability distribution model to calculate the wind and solar power output sample set; the optimization objective balances the minimization of Euclidean distance between sample points and the uniformity of projection of each variable dimension through weighting coefficients. Step 3: Traverse the sample points in the wind and solar power output sample set, and calculate the system carbon flow rate and its probability distribution characteristics based on the system carbon flow rate comprehensive analysis model; Step 4: Based on the reverse current method and the path output distribution factor theory, for each sample point in the wind and solar power output sample set, calculate the changes in system node carbon flow rate, branch carbon flow rate, and network loss carbon flow rate caused by wind and solar power output injection; and perform a weighted average to obtain the average influence factor of wind and solar power output on system node carbon flow rate, branch carbon flow rate, and network loss carbon flow rate.

[0008] More preferably, The comprehensive analysis model of system carbon flow rate under the uncertainty of wind and solar power output coupling is used to quantify the mapping relationship between the wind and solar power output with random fluctuations and the total carbon flow rate of the system, specifically:

[0009]

[0010] in, The total carbon flow rate of the system. For the combined wind and solar power output vector, It is a unit vector; For the first s The carbon emission intensity of the node, and the first node sEach node is a node for the balanced generator unit access; is a constant, representing all deterministic factors other than wind and light fluctuations; For the first, excluding the balancing unit The generator injection power of each node, with this value being 0 for load nodes; For the first The load power of each node; , They are nodes j ,node m Load power, node j Indicates the wind farm access node, node m Indicates the grid connection node of a photovoltaic power station; This is due to network loss in the system. , The first k Carbon emission intensity and active power of each node connected to the unit; The combined wind and solar power output vector Specifically: ,in, For the output power of the wind farm, Wind speed; For the output power of photovoltaic power plants, Solar irradiance.

[0011] More preferably, The total carbon flow rate of the system Let be a random variable driven by the combined output of wind and solar power, with its expected value and variance as follows:

[0012]

[0013] in, Let be the mathematical expectation of the total carbon flow rate of the system; Let V be the variance of the total carbon flow rate of the system; , These are the mathematical expectations of the output power of the wind farm and the output power of the photovoltaic power station, respectively. , These are the variances of the output power of the wind farm and the output power of the photovoltaic power station, respectively. Covariance contributing to wind and solar power.

[0014] More preferably, In step 2, the objective function for maximizing spatial dispersion is as follows:

[0015] in, It represents a sequence of sample subsets, that is, a permutation and combination; , They are respectively i The, the j One sample point; The Euclidean distance between sample points; The total number of all sample points; D is the total number of dimensions of the variables; The number of intervals divided for each dimension variable; For the first The dimension variable falls on the th The number of samples in each interval; , , which are weighting coefficients used to balance the two objectives of distance between points and projection uniformity.

[0016] More preferably, Step 2, the adaptive sequence search algorithm, specifically includes the following steps: Step 2.1: Randomly generate an initial sample sequence and calculate its objective function value; Step 2.2: Define the neighborhood operation as swapping the positions of any two sample subsets in the sample sequence, and then perform an iterative search, specifically: Step 2.2.1: After performing the neighborhood operation on the current sample sequence, a new sample sequence is generated. If the objective function value of the new sample sequence is better, it is directly accepted as the current optimal solution. Otherwise, the probability of acceptance is calculated according to the Metropolis criterion. P The probability is combined with a random number. In comparison, if If the objective function value temporarily deteriorates, then the new sample sequence that causes it to deteriorate is accepted as the current optimal solution to avoid getting trapped in a local optimum; if Then the original sample sequence is taken as the current optimal solution, where, satisfy ; Step 2.2.2: Using the sample sequence corresponding to the current optimal solution as a benchmark, perform a fine search only on the neighborhood sample sequences that can be generated through adjacent subset swapping and subset position sliding operations, until continuous m The objective function value was not improved in the second instance, where, m This is the threshold for the number of iterations; Step 2.2.3: If continuous N glo If a better solution is not found in the next iteration, the perturbation amplitude is increased to enhance the global exploration capability; if the near... N loc If the objective function value increases continuously in each iteration, then the perturbation amplitude is reduced, and a finer search is performed; where... N gloThe threshold for triggering global exploration, N loc This is the threshold for triggering a local fine-tuning search. Step 2.3: When the preset maximum number of iterations is reached... C limit Or the objective function value is in continuous L The iteration terminates when the relative improvement is less than a preset relative threshold; where the relative improvement is the difference between the objective function value of the current iteration and the objective function value of the previous iteration divided by the absolute value of the objective function value of the previous iteration.

[0017] More preferably, The acceptance probability is calculated based on the Metropolis criterion. P The calculation method is as follows:

[0018] in, The objective function value for the current sample sequence; The objective function value of the new sample sequence after perturbation; For the first k The control parameters at the time of the next iteration, and as the number of iterations increases, According to the attenuation coefficient Gradually decrease, that is , satisfy ; In step 2.2.3, increasing the disturbance amplitude specifically includes increasing the temperature attenuation coefficient. The reduction of disturbance amplitude includes reducing the temperature attenuation coefficient. .

[0019] More preferably, In step 4, the changes in system node carbon flow rate, branch carbon flow rate, and network loss carbon flow rate caused by the wind and solar power output injection are calculated as follows:

[0020]

[0021]

[0022] in, , These represent the changes in output power of wind farms and output power of photovoltaic power plants, respectively. , , These represent the changes in carbon flow rate at system nodes, carbon flow rate at branch lines, and carbon flow rate at network losses, respectively. To balance the influence factor vector of node carbon flow on the unit, To balance the influence factor vector of the unit on the branch carbon flow, The vector of factors affecting the carbon flow of the balancing unit on network losses.

[0023] More preferably, The influence factor vector of the balancing unit on the node carbon flow is determined in the following manner:

[0024] in, To balance the carbon emission intensity of the unit; To represent the balance node s Indicator vector, Output the distribution matrix for the carbon flow path; The influence factor vector of the balancing unit on the branch carbon flow is determined in the following manner:

[0025] in, The node output distribution matrix for carbon flow; The influence factor vector of the balancing unit on network loss carbon flow is determined in the following manner:

[0026] in, It is a row vector consisting entirely of 1s. This is the active power loss matrix for the branch. This is the system node flux matrix.

[0027] More preferably, In step 4, the average influence factor of the wind and solar power output on the carbon flow rate of system nodes, branch carbon flow rate, and network loss carbon flow rate is determined as follows: The average impact factors of wind power output on system node carbon flow rate, branch carbon flow rate, and network loss carbon flow rate are as follows:

[0028]

[0029]

[0030] in, , , The hierarchical Latin hypercube sampling method was used for the following purposes: The average impact factor of wind power output on system node carbon flow rate, branch carbon flow rate and network loss carbon flow rate after round sampling; K For the total number of sampling rounds; , , The first k The vectors of the balancing unit's influence on node carbon flow, the vectors of the balancing unit's influence on branch carbon flow, and the vectors of the balancing unit's influence on network loss carbon flow are corresponding to the subsamples. , The first k The changes in the output power of wind farms and photovoltaic power plants corresponding to the sub-samples; The average impact factors of photovoltaic output on the carbon flow rate of system nodes, branch carbon flow rate, and network loss carbon flow rate are as follows:

[0031]

[0032]

[0033] in, , , They are respectively The average impact factor of photovoltaic output on system node carbon flow rate, branch carbon flow rate and network loss carbon flow rate after round sampling.

[0034] Another aspect of the present invention discloses a wind-solar carbon flow correlation analysis system based on PLHS sampling and sample optimization, which includes a system carbon flow rate comprehensive analysis model construction module, a wind and solar power output sample set acquisition module, a system carbon flow rate calculation module, and a wind and solar injection carbon flow rate impact analysis module. The module for constructing a comprehensive analysis model of system carbon flow rate is used to build a probability distribution model of wind and solar power output and a comprehensive analysis model of system carbon flow rate under the coupling uncertainty of wind and solar power output. The wind and solar power output sample set acquisition module uses a hierarchical Latin hypercube sampling method to sample wind speed and light intensity to obtain a sample sequence. An adaptive sequence search algorithm is proposed, which uses a composite index based on spatial dispersion as the optimization objective to optimize the sample sequence. The optimized sample sequence is then input into the wind and solar power output probability distribution model to calculate the wind and solar power output sample set. The optimization objective balances the minimization of Euclidean distance between sample points and the uniformity of the projection of each variable dimension through weighting coefficients. The system carbon flow rate calculation module traverses the sample points in the wind and solar power output sample set and calculates the system carbon flow rate based on the system carbon flow rate comprehensive analysis model. The wind and solar power injection carbon flow rate impact analysis module, based on the reverse tidal current method and path output distribution factor theory, calculates the changes in system node carbon flow rate, branch carbon flow rate, and network loss carbon flow rate caused by wind and solar power injection for each sample point in the wind and solar power output sample set; and performs a weighted average to obtain the average impact factor of wind and solar power on system node carbon flow rate, branch carbon flow rate, and network loss carbon flow rate.

[0035] The beneficial effects of this invention are that, compared with the prior art, This invention constructs a complete probabilistic carbon flow analysis framework for power systems that considers wind and solar uncertainties, and proposes an efficient calculation method based on hierarchical Latin hypercube sampling, achieving accurate assessment of the carbon emission distribution characteristics of the system. Through theoretical analysis and numerical examples, the following conclusions are obtained: 1) The progressive hierarchical sampling strategy adopted in this invention significantly improves computational efficiency while ensuring sampling accuracy. It can achieve estimation results similar to traditional Monte Carlo methods with fewer samples, effectively solving the computational efficiency bottleneck problem in high-dimensional uncertainty analysis.

[0036] 2) Compared with traditional deterministic carbon flow analysis methods, the method proposed in this invention can fully characterize the impact of the randomness of wind and solar power output on the carbon flow distribution of the system. It can not only provide the expected value of carbon flow rate, but also quantify its fluctuation range and probability distribution characteristics, providing a more comprehensive decision-making basis for the low-carbon operation of the system.

[0037] 3) While simplified models can reflect the carbon flow characteristics of the system to some extent, they are difficult to accurately capture the complex coupling relationship between wind and solar uncertainties and system carbon flow. The refined probabilistic carbon flow model established in this invention, by comprehensively considering the correlation between wind and solar power output, network topology, and unit carbon emission characteristics, can more accurately assess the impact of new energy access on the system's carbon footprint.

[0038] For a detailed explanation of the advantages of this invention, please refer to the embodiment analysis section of this invention. Attached Figure Description

[0039] Figure 1 This is a flowchart of the probabilistic power flow analysis based on PLHS sampling and sample sequence optimization in this invention; Figure 2 This is a complete flowchart of the probabilistic carbon flow calculation based on PLSH in Embodiment 1 of the present invention; Figure 3 This is the IEEE 14-node test system in Embodiment 2 of the present invention; Figure 4 This is the active power probability density curve of branch 12-13 in Embodiment 2 of the present invention; Figure 5 This is the cumulative active power distribution curve of branches 12-13 in Embodiment 2 of the present invention; Figure 6 This is the cumulative probability distribution curve of the total network loss in the system in Embodiment 2 of the present invention; Figure 7 This is the expected relative error of the voltage amplitude at node 12 in Embodiment 2 of the present invention; Figure 8 This refers to the relative variance error of the voltage amplitude at node 12 in Embodiment 2 of the present invention. Figure 9 This refers to random wind speed sampling in Embodiment 3 of the present invention; Figure 10 This is a graph showing the relationship between wind speed and total carbon flow rate in Embodiment 3 of the present invention; Figure 11 This refers to random light intensity sampling in Embodiment 3 of the present invention; Figure 12 This is a graph showing the relationship between light intensity and total carbon flow rate of the system in Embodiment 3 of the present invention; Figure 13 This is a graph showing the relationship between the combined effect of wind and solar power and the total carbon flow rate of the system in Embodiment 3 of the present invention. Figure 14 This refers to the branch carbon flow rate in each scenario of Embodiment 3 of the present invention; Figure 15 It is the sum of the branch carbon flow rates in each scenario in Embodiment 3 of the present invention; Figure 16 These are the node carbon potentials in various scenarios in Embodiment 3 of the present invention; Figure 17 It is the average node carbon potential in each scenario in Embodiment 3 of the present invention. Detailed Implementation

[0040] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of this invention. The embodiments described in this application are merely some embodiments of this invention, and not all embodiments. Based on the spirit of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of this invention.

[0041] like Figure 1 As shown, this invention discloses a wind-solar-carbon flow correlation analysis method based on PLHS sampling and sample optimization, characterized by the following steps: Step 1: Construct a probability distribution model for wind and solar power output; construct a comprehensive analysis model for system carbon flow rate under the coupling uncertainty of wind and solar power output; The comprehensive analysis model of system carbon flow rate under the uncertainty of wind and solar power output coupling is used to quantify the mapping relationship between the wind and solar power output with random fluctuations and the total carbon flow rate of the system, specifically:

[0042]

[0043] in, The total carbon flow rate of the system. For the combined wind and solar power output vector, It is a unit vector; For the first s The carbon emission intensity of the node, and the first node s Each node is a node for the balanced generator unit access; is a constant, representing all deterministic factors other than wind and light fluctuations; For the first, excluding the balancing unit The generator injection power of each node, with this value being 0 for load nodes; For the first The load power of each node; , They are nodes j ,node m Load power, node j Indicates the wind farm access node, node m Indicates the grid connection node of a photovoltaic power station; This is due to network loss in the system. , The first k Carbon emission intensity and active power of each node connected to the unit; The combined wind and solar power output vector Specifically: ,in, For the output power of the wind farm, Wind speed; For the output power of photovoltaic power plants, Solar irradiance.

[0044] The total carbon flow rate of the system Let be a random variable driven by the combined output of wind and solar power, with its expected value and variance as follows:

[0045]

[0046] in, Let be the mathematical expectation of the total carbon flow rate of the system; Let V be the variance of the total carbon flow rate of the system; , These are the mathematical expectations of the output power of the wind farm and the output power of the photovoltaic power station, respectively. , These are the variances of the output power of the wind farm and the output power of the photovoltaic power station, respectively. Covariance contributing to wind and solar power.

[0047] Step 2: A hierarchical Latin hypercube sampling method is used to sample wind speed and light intensity to obtain a sample sequence; an adaptive sequence search algorithm is proposed, which uses a composite index based on spatial dispersion as the optimization objective to optimize the sample sequence. The optimized sample sequence is then input into the wind and solar power output probability distribution model to calculate the wind and solar power output sample set; the optimization objective balances the minimization of Euclidean distance between sample points and the uniformity of projection of each variable dimension through weighting coefficients. The objective function for maximizing spatial dispersion is as follows:

[0048] in, It represents a sequence of sample subsets, that is, a permutation and combination; , They are respectively i The, the j One sample point; The Euclidean distance between sample points; The total number of all sample points; D is the total number of dimensions of the variables; The number of intervals divided for each dimension variable; For the first The dimension variable falls on the th The number of samples in each interval; , , which are weighting coefficients used to balance the two objectives of distance between points and projection uniformity.

[0049] Step 2, the adaptive sequence search algorithm, specifically includes the following steps: Step 2.1: Randomly generate an initial sample sequence and calculate its objective function value; Step 2.2: Define the neighborhood operation as swapping the positions of any two sample subsets in the sample sequence, and then perform an iterative search, specifically: Step 2.2.1: After performing the neighborhood operation on the current sample sequence, a new sample sequence is generated. If the objective function value of the new sample sequence is better, it is directly accepted as the current optimal solution. Otherwise, the probability of acceptance is calculated according to the Metropolis criterion. P The probability is combined with a random number. In comparison, if If the objective function value temporarily deteriorates, then the new sample sequence that causes it to deteriorate is accepted as the current optimal solution to avoid getting trapped in a local optimum; if Then the original sample sequence is taken as the current optimal solution, where, satisfy ; Step 2.2.2: Using the sample sequence corresponding to the current optimal solution as a benchmark, perform a fine search only on the neighborhood sample sequences that can be generated through adjacent subset swapping and subset position sliding operations, until continuous m The objective function value was not improved in the second instance, where, m This is the threshold for the number of iterations; Step 2.2.3: If continuous N glo If a better solution is not found in the next iteration, the perturbation amplitude is increased to enhance the global exploration capability; if the near... N loc If the objective function value increases continuously in each iteration, then the perturbation amplitude is reduced, and a finer search is performed; where... N glo The threshold for triggering global exploration, N loc This is the threshold for triggering a local fine-tuning search. Step 2.3: When the preset maximum number of iterations is reached... C limit Or the objective function value is in continuous L The iteration terminates when the relative improvement is less than a preset relative threshold; where the relative improvement is the difference between the objective function value of the current iteration and the objective function value of the previous iteration divided by the absolute value of the objective function value of the previous iteration.

[0050] The acceptance probability is calculated based on the Metropolis criterion. P The calculation method is as follows:

[0051] in, The objective function value for the current sample sequence; The objective function value of the new sample sequence after perturbation; For the first k The control parameters at the time of the next iteration, and as the number of iterations increases, According to the attenuation coefficient Gradually decrease, that is , satisfy ; In step 2.2.3, increasing the disturbance amplitude specifically includes increasing the temperature attenuation coefficient. The reduction of disturbance amplitude includes reducing the temperature attenuation coefficient. .

[0052] Step 3: Traverse the sample points in the wind and solar power output sample set, and calculate the system carbon flow rate and its probability distribution characteristics based on the system carbon flow rate comprehensive analysis model; Step 4: Based on the reverse current method and the path output distribution factor theory, for each sample point in the wind and solar power output sample set, calculate the changes in system node carbon flow rate, branch carbon flow rate, and network loss carbon flow rate caused by wind and solar power output injection; and perform a weighted average to obtain the average influence factor of wind and solar power output on system node carbon flow rate, branch carbon flow rate, and network loss carbon flow rate.

[0053] The changes in system node carbon flow rate, branch carbon flow rate, and network loss carbon flow rate caused by the wind and solar power output injection are calculated in the following manner:

[0054]

[0055]

[0056] in, , These represent the changes in output power of wind farms and output power of photovoltaic power plants, respectively. , , These represent the changes in carbon flow rate at system nodes, carbon flow rate at branch lines, and carbon flow rate at network losses, respectively. To balance the influence factor vector of node carbon flow on the unit, To balance the influence factor vector of the unit on the branch carbon flow, The vector of factors affecting the carbon flow of the balancing unit on network losses.

[0057] The influence factor vector of the balancing unit on the node carbon flow is determined in the following manner:

[0058] in, To balance the carbon emission intensity of the unit; To represent the balance node s Indicator vector, Output the distribution matrix for the carbon flow path; The influence factor vector of the balancing unit on the branch carbon flow is determined in the following manner:

[0059] in, The node output distribution matrix for carbon flow; The influence factor vector of the balancing unit on network loss carbon flow is determined in the following manner:

[0060] in, It is a row vector consisting entirely of 1s. This is the active power loss matrix for the branch. This is the system node flux matrix.

[0061] The average impact factor of wind and solar power output on system node carbon flow rate, branch carbon flow rate, and network loss carbon flow rate is determined as follows: The average impact factors of wind power output on system node carbon flow rate, branch carbon flow rate, and network loss carbon flow rate are as follows:

[0062]

[0063]

[0064] in, , , The hierarchical Latin hypercube sampling method was used for the following purposes: The average impact factor of wind power output on system node carbon flow rate, branch carbon flow rate and network loss carbon flow rate after round sampling; K For the total number of sampling rounds; , , The first k The vectors of the balancing unit's influence on node carbon flow, the vectors of the balancing unit's influence on branch carbon flow, and the vectors of the balancing unit's influence on network loss carbon flow are corresponding to the subsamples. , The first k The changes in the output power of wind farms and photovoltaic power plants corresponding to the sub-samples; The average impact factors of photovoltaic output on the carbon flow rate of system nodes, branch carbon flow rate, and network loss carbon flow rate are as follows:

[0065]

[0066]

[0067] in, , , They are respectively The average impact factor of photovoltaic output on system node carbon flow rate, branch carbon flow rate and network loss carbon flow rate after round sampling.

[0068] The present invention also discloses a wind-solar carbon flow correlation analysis system based on PLHS sampling and sample optimization based on the aforementioned method, including a system carbon flow rate comprehensive analysis model construction module, a wind and solar power output sample set acquisition module, a system carbon flow rate calculation module, and a wind and solar injection carbon flow rate impact analysis module. The module for constructing a comprehensive analysis model of system carbon flow rate is used to build a probability distribution model of wind and solar power output and a comprehensive analysis model of system carbon flow rate under the coupling uncertainty of wind and solar power output. The wind and solar power output sample set acquisition module uses a hierarchical Latin hypercube sampling method to sample wind speed and light intensity to obtain a sample sequence. An adaptive sequence search algorithm is proposed, which uses a composite index based on spatial dispersion as the optimization objective to optimize the sample sequence. The optimized sample sequence is then input into the wind and solar power output probability distribution model to calculate the wind and solar power output sample set. The optimization objective balances the minimization of Euclidean distance between sample points and the uniformity of the projection of each variable dimension through weighting coefficients. The system carbon flow rate calculation module traverses the sample points in the wind and solar power output sample set and calculates the system carbon flow rate based on the system carbon flow rate comprehensive analysis model. The wind and solar power injection carbon flow rate impact analysis module, based on the reverse tidal current method and path output distribution factor theory, calculates the changes in system node carbon flow rate, branch carbon flow rate, and network loss carbon flow rate caused by wind and solar power injection for each sample point in the wind and solar power output sample set; and performs a weighted average to obtain the average impact factor of wind and solar power on system node carbon flow rate, branch carbon flow rate, and network loss carbon flow rate.

[0069] Example 1: like Figure 1 As shown, this invention proposes a wind-solar-carbon flow correlation analysis method based on PLHS sampling and sample optimization, specifically including the following steps: Step 1: Construct a probability distribution model for wind and solar power output; construct a comprehensive analysis model for system carbon flow rate under the coupling uncertainty of wind and solar power output; The uncertainty in calculating carbon flow in the power system mainly stems from the random fluctuations in renewable energy output. To fully quantify its impact on carbon footprint distribution, this invention combines mainstream probabilistic carbon flow analysis methods and selects the joint uncertainty of wind farm output and photovoltaic power plant output for modeling; specifically, it includes the following steps: (1) System carbon flow rate modeling for the stochasticity of wind farm output 1) Wind turbine power characteristic model The output power of a wind farm is influenced by a combination of factors, with wind speed being the primary one. Generally, wind speed in a wind farm does not follow a normal distribution but often a Weibull distribution. Ignoring power losses during wind power generation, the active power output of the wind power injection system can be considered as the mechanical power of all the wind turbines in the wind farm. Considering a single turbine, the wind speed... And wind turbine output power The functional relationship can be described by the following function:

[0070] In the formula: This refers to the rated output power of the wind turbine. To cut into wind speed, Rated wind speed, To cut off the wind speed.

[0071] For including The total output power of a wind farm with the same type of wind turbine. for:

[0072] 2) Wind speed uncertainty modeling Weibull distribution probability density function model of active power of wind turbine:

[0073] In the formula: , All are Weibull distribution parameters; For shape parameters; For scale parameters; This represents the current actual wind speed.

[0074] 3) System power balance Assuming the wind farm access node is The balance node is Output of the balancing unit It is necessary to compensate for the power output fluctuations of wind farms and system network losses, namely:

[0075] In the formula: The first one besides the balancing unit The generator injection power of the node; this value is 0 for load nodes. For the first The load power of each node; This is the sum of the total system losses; N This represents the total number of nodes.

[0076] 4) Carbon Flow Rate Correlation Model of Wind Power System For the system as a whole, the total carbon emissions per unit time are the sum of the carbon flow rates of each node. It is equal to the sum of the carbon emissions produced by each unit per unit time:

[0077] In the formula: Indicates the number of conventional generating units; , For the first k Carbon emission intensity and active power of each node connected to the unit; , These are the carbon emission intensity and active power of the balancing unit, respectively.

[0078] Substituting the output expression of the balancing unit into the above formula, the total carbon flow rate of the system after the wind farm is connected can be established. With wind speed Explicit functional relationships:

[0079] The model clearly shows that the randomness of wind speed affects the power output of wind farms. Fluctuations, in turn, affect the output of the balancing unit through the power balance relationship, ultimately impacting the total carbon flow rate of the system. It becomes a random variable that varies with wind speed.

[0080] (2) System carbon flow rate modeling for the stochasticity of photovoltaic output The fluctuation in the output power of photovoltaic power plants mainly stems from the random variation of solar irradiance. This section aims to establish a quantitative relationship model between irradiance and the total carbon flow rate of the system, providing a theoretical basis for constructing a probabilistic carbon flow analysis framework.

[0081] 1) Photovoltaic array output power model The output power of a photovoltaic array is mainly affected by solar irradiance and ambient temperature, and its mathematical expression can be described as follows:

[0082] In the formula: This represents the total active power output of the photovoltaic array. This represents the actual solar irradiance received by the inclined surface of the photovoltaic panel. The ambient temperature; This represents the number of modules connected in series in the photovoltaic array; The effective light-receiving area of ​​a single photovoltaic module; The photoelectric conversion efficiency of photovoltaic modules under standard test conditions (STC); The power temperature coefficient of a photovoltaic module; This is the standard test temperature.

[0083] To focus on the core impact of irradiance uncertainty, the above model can be reasonably simplified. Ignoring the second-order effects of temperature changes, the photovoltaic output power can be... Simplified to a linear function of irradiance:

[0084] In the formula: Let be the power conversion coefficient of the photovoltaic power plant, and .

[0085] 2) Probability distribution model of light intensity Solar irradiance It exhibits significant randomness, and its probability distribution characteristics are key to characterizing the uncertainty of photovoltaic power output. Studies have shown that irradiance within a specific time period... It follows a Beta probability distribution very well, and its probability density function (PDF) is:

[0086] In the formula: The maximum possible irradiance for the given geographical location during that time period; For Beta functions, , The shape parameter of the Beta distribution can be obtained from the mean of historical irradiance data. and variance Make an estimate:

[0087]

[0088] In the formula: and The mean and standard deviation of irradiance after normalization to the interval [0,1].

[0089] 3) Total carbon flow rate correlation model of photovoltaic system Assuming the photovoltaic power station is connected to the system node The system balance node is According to the power balance principle, the output of the balancing unit... The compensation for photovoltaic output fluctuations, load fluctuations, and system network losses is expressed as follows:

[0090] In the formula: The first one besides the balancing unit The generator injection power of the node; this value is 0 for load nodes. For the first The load power of each node; For the first The load power of each node; This is due to network loss in the system.

[0091] Total carbon flow rate of the system Carbon emissions from all generators can be expressed as:

[0092] In the formula: Indicates the number of conventional generating units; , For the first k Carbon emission intensity and active power of each node connected to the unit; , These are the carbon emission intensity and active power of the balancing unit, respectively.

[0093] Output of the balancing unit Substituting the expression into the above equation, we can establish the total carbon flow rate of the system. With solar irradiance Direct functional relationship:

[0094] in, The constant term characterizing the influence of all fixed and deterministic factors within the system, excluding photovoltaic fluctuations, is expressed as follows:

[0095] (3) Comprehensive analysis model of system carbon flow rate under the uncertainty of wind and solar power output coupling When wind farms and photovoltaic power plants are connected to the grid simultaneously, the randomness and intermittency of their power output are coupled, jointly affecting the system power balance and having a combined impact on the total carbon flow rate. This section aims to construct a unified mathematical model to characterize the mapping relationship between the coordinated fluctuations of wind and solar resources and the total carbon emissions of the system.

[0096] 1) Wind-solar combined output vector representation To comprehensively analyze the uncertainties in wind and solar power output, a combined wind and solar power output vector is defined. as follows:

[0097] in, The output power of the wind farm is a function of wind speed; The output power of a photovoltaic power station is solar irradiance. The function of this vector comprehensively represents the injected power state of renewable energy.

[0098] 2) General expression for system power balance and carbon flow rate Set up wind farm access node Photovoltaic power plant grid connection node The system balance node is Based on the principle of system power balance, the output of the balancing unit is... for:

[0099] in, It is a unit vector. This refers to the combined efforts of wind and solar power. , They are nodes j ,node m The load power.

[0100] Total carbon flow rate of the system The sum of carbon emissions from all units:

[0101] Will Substituting into the above equation, we obtain the general expression for the total carbon flow rate considering combined wind and solar power output:

[0102] Among them, the constant term It integrates all deterministic factors that are not related to wind and light fluctuations:

[0103] 3) Probabilistic carbon flow rate model under uncertainties of wind-solar integration Due to wind speed and irradiance All are random variables and may have some correlation; joint output vector It is also a random vector. Therefore, the total carbon flux of the system... It becomes a random variable jointly driven by wind and solar resources.

[0104] Its mathematical expectation It can be represented as:

[0105] In the formula: for The mathematical expectation; for The mathematical expectation.

[0106] variance This reflects the range of carbon flow rate fluctuations caused by uncertainties in the wind-solar integration process:

[0107] In the formula, The covariance of wind and solar power outputs characterizes their spatiotemporal correlation. Positive correlation exacerbates fluctuations in the system's total carbon flux, while negative correlation can mitigate these fluctuations to some extent.

[0108] Step 2: A hierarchical Latin hypercube sampling method is used to sample wind speed and light intensity to obtain a sample sequence; an adaptive sequence search algorithm is proposed, which uses a composite index based on spatial dispersion as the optimization objective to optimize the sample sequence. The optimized sample sequence is then input into the wind and solar power output probability distribution model to calculate the wind and solar power output sample set; the optimization objective balances the minimization of Euclidean distance between sample points and the uniformity of projection of each variable dimension through weighting coefficients. The core of the probabilistic carbon flow analysis framework of this invention lies in the adoption of an efficient sampling strategy—Progressive Latin Hypercube Sampling (PLHS)—to drive large-scale deterministic carbon flow calculations, thereby converging the probability distribution of the system's carbon flow indices.

[0109] The mathematical principle of the hierarchical Latin hypercube sampling method and its specific implementation process in this invention are as follows: 1. A hierarchical Latin hypercube sampling method was used to sample wind speed and light intensity to obtain sample sequences; specifically including: (1) Preparation of basic sampling data First, determine all random input variables (such as wind speed). Light intensity The probability distribution model and parameters of the power flow are determined. A deterministic power flow calculation model is established. ,in For the input variable vector, This is the system state output vector (such as node voltage and branch power flow).

[0110] (2) Structured progressive sampling strategy PLHS is an advanced Monte Carlo method designed to improve sampling efficiency and space-filling performance. Its core idea is to streamline a large-scale sampling task. (in The total number of samples, The random variable dimension is decomposed into a series of ordered, smaller sampling stages, denoted as... , , ..., ;in, This represents the total number of stages. The number of samples in each stage satisfies .

[0111] (3) Basic Mathematical Principles of PLSH Phase-wise Inner Latin Hypercube Properties: Each sampling phase Each of these is itself a complete Latin hypercube sample. This means that for each dimension of the random variable, its... The sample values ​​are exactly uniformly distributed in Within each equally probable interval, the uniformity of the one-dimensional projection within the stage is guaranteed. Mathematically, if the... The cumulative distribution function of the dimensional random variable is Then the stage Sample values ​​in this dimension satisfy:

[0112] And all sample values ​​are randomly distributed within this interval.

[0113] Inter-stage asymptotic unbiasedness: Different sampling stages are generated through a coordination mechanism to ensure that early stages (such as...) The sample distribution will not affect subsequent stages (such as...). The samples contain structural biases. This allows the sampling process to be interrupted at any time, affecting the currently acquired cumulative samples. It remains an unbiased representation of the entire parameter space.

[0114] Global space filling: By optimizing the permutation and combination of sample sets at each stage, PLHS strives to make the final merged global sample set as perfect as possible. exist The sample points are distributed as evenly as possible in the dimensional space to maximize the minimum distance between them, thus outperforming the space-filling characteristics of simple random sampling or even traditional LHS.

[0115] (4) Generation of hierarchical Latin hypercube samples First, a uniform distribution transformation is performed, based on the probability distribution of each random variable. The cumulative distribution function is used to uniformly transform the samples to a uniform distribution in the [0,1] interval. Then, dimensional stratification and slicing are performed to generate the total sample size. Divided into Each stratum contains [number] levels. One sample. Within each level, at... A Latin hypercube sample slice is generated independently in the 1-dimensional unit hypercube space. This ensures that each slice satisfies the Latin hypercube property within its interior.

[0116] 2. An adaptive sequence search algorithm is proposed, which uses a composite index based on spatial dispersion as the optimization objective to optimize the sample sequence; Specifically, the collaborative optimization recombination process for PLSH-sampled sequence sequences is as follows: In this step, to improve the spatial representation capability of the overall sample set, the initially generated sample subsets need to be sequentially optimized and recombined. This process involves constructing a collaborative optimization model to find the optimal permutation and combination that maximizes the spatial distribution characteristics of the merged global sample set.

[0117] (1) Optimize the construction of the objective function Define an index function to evaluate the quality of the spatial distribution of a sample set. This function should effectively reflect the dispersion of sample points in a multidimensional space. This invention uses a composite index based on Spatial Dispersion Index (SDI) as the optimization objective:

[0118] in, It represents a sequence of sample subsets, that is, a permutation and combination; , They are respectively i The, the j 1000 sample points (each sample point is a vector containing D-dimensional variables); The Euclidean distance between sample points; The total number of all sample points; D is the total number of dimensions of the variables; The number of intervals divided for each dimension variable; For the first The dimension variable falls on the th The number of samples in each interval; , These are weighting coefficients used to balance the two objectives of distance between points and projection uniformity, and satisfy the following conditions: .

[0119] The optimization goal is to find The largest subset sequence of samples.

[0120] (2) Adaptive Sequence Search Algorithm To solve the above optimization problem, an Adaptive Sequence Search (ASS) algorithm is designed. This algorithm does not rely on a single greedy strategy, but combines global exploration and local fine-grained search. The adaptive sequence search algorithm specifically includes the following steps: 1) Initialization: Randomly generate an initial sample sequence And calculate its objective function value. .

[0121] 2) Neighborhood operation: The neighborhood operation is defined as swapping the positions of any two sample subsets in the sample sequence.

[0122] 3) Iterative search: a) Perturbation phase: For the current sequence Perform neighborhood operations to obtain a new sample sequence. : If the new sample sequence Corresponding objective function value Then accept directly. As the current optimal solution; If the new sample sequence Corresponding objective function value Then, the probability of acceptance is calculated according to the Metropolis criterion. P The calculation formula is as follows:

[0123] in, The objective function value for the current sample sequence; The objective function value of the new sample sequence after perturbation; For the first k The control parameters at the time of the next iteration, and as the number of iterations increases, According to the attenuation coefficient Gradually decrease, that is In this embodiment, preferably, ; Combine the probability with random numbers In comparison, if Then accept the neighboring new sample sequence that temporarily worsens the objective function value. As the current optimal solution, to avoid getting trapped in local optima; if If the new solution fails, the original sample sequence is rejected and retained as the current optimal solution; where, satisfy ; b) Reinforcement Phase: Using the sample sequence corresponding to the current optimal solution as a benchmark, a fine search is performed only on the neighborhood sample sequences that can be generated through adjacent subset swapping and subset position sliding operations, until a continuous... m The objective function value was not improved in the second instance, where, m This is the threshold for the number of iterations, typically ranging from 10 to 20. c) Adaptive Adjustment: Dynamically adjusts the probability and magnitude of the "perturbation" based on the search progress. If continuous... N glo If a better solution is not found in the next iteration, the perturbation amplitude is increased to enhance the global exploration capability; if the near... N loc If the objective function value increases continuously in each iteration, then the perturbation amplitude is reduced, and a finer search is performed; where... N glo This is the threshold for triggering global exploration, with a typical value of 20-30. N loc This is the threshold for triggering local fine-grained searches, with a typical value of 10~15; The increase in disturbance amplitude includes increasing the temperature attenuation coefficient. Expand the scope of domain operations, such as allowing exchange distances greater than d. th A subset of; The reduction of disturbance amplitude includes reducing the temperature attenuation coefficient. Narrow the scope of neighborhood operations, such as allowing only adjacent subsets to be swapped; 4) Termination condition: The algorithm terminates when it reaches the preset maximum number of iterations. C limit Or the objective function value is in continuous L The iteration terminates when the relative improvement is less than a preset relative threshold. The relative improvement is the difference between the objective function value of the current iteration and the objective function value of the previous iteration, divided by the absolute value of the objective function value of the previous iteration. The preset relative threshold is typically 0.1%.

[0124] (3) Generation of the optimal sample set The ASS algorithm ultimately outputs an optimized subset sequence of samples. All sample subsets are merged according to this optimal sequence to obtain the final global sample set. :

[0125] in, To optimize the first in the sequence A subset of samples; this sample set Not only does it exhibit good edge distribution characteristics in each dimension, but more importantly, the uniformity of its sample point distribution and spatial filling in the multidimensional joint space are significantly enhanced, providing high-quality input samples for subsequent probabilistic carbon flow calculations. Figure 1 The specific analysis process is demonstrated.

[0126] Step 3: Traverse the sample points in the wind and solar power output sample set, and calculate the system carbon flow rate and its probability distribution characteristics based on the system carbon flow rate comprehensive analysis model; Specifically, by directly substituting the samples generated in step 2 into the system carbon flow rate calculation formula and its corresponding expectation and variance formulas in step 1, the system carbon flow rate and its probability distribution characteristics can be obtained. Furthermore, based on the calculated nodal carbon flux, the nodal carbon potential can be calculated. ,in For nodes carbon flow rate, For nodes The active power.

[0127] Step 4: Based on the reverse current method and the path output distribution factor theory, for each sample point in the wind and solar power output sample set, calculate the changes in system node carbon flow rate, branch carbon flow rate, and network loss carbon flow rate caused by wind and solar power output injection; and perform a weighted average to obtain the average influence factor of wind and solar power output on system node carbon flow rate, branch carbon flow rate, and network loss carbon flow rate.

[0128] In step 3, this invention has obtained the probability distribution characteristics (mathematical expectation and variance) of the total carbon flow rate of the system under wind-solar combined drive, quantifying the overall low-carbon operation level and fluctuation risk of the system. To further analyze the spatial distribution of the overall fluctuations in the power grid topology and identify the key nodes and paths leading to increased system carbon flow rate variance, it is necessary to establish an average influence factor model for the carbon flow rate of system nodes, branches, and network losses caused by wind and solar injected power. This section, based on the countercurrent power flow method and path output distribution factor theory, constructs a method for calculating the carbon flow influence factor under wind-solar combined uncertainty, deconstructing the system-level uncertainty index in step 3 to the node and branch levels.

[0129] To accurately assess the carbon flow distribution characteristics of a system under wind-solar stochastic conditions, it is necessary to establish an average influence factor model for the carbon flow rate of system nodes, branches, and network losses caused by wind and solar power injection. This section constructs a method for calculating the carbon flow influence factor under combined wind and solar uncertainties based on the countercurrent tidal method and path output distribution factor theory.

[0130] (1) Calculation of carbon flow distribution correlation matrix The core characterization of carbon flow distribution relies on two key matrices: the node output distribution matrix. With path output distribution matrix The node output distribution matrix is ​​shown below. It is mainly used to depict the distribution of active power flow and carbon flow within a power system under steady-state operation scenarios. The elements of this matrix accurately reflect the proportion of branch carbon flow flowing from a given node to its directly connected nodes, relative to the total network flow flowing into that node.

[0131] The path output distribution matrix This focuses on describing the active power flow and carbon flow injected into the system by generator sets, and the path information from the generator sets to each node and load in the system. The elements of this matrix reflect the contribution of the carbon flow flowing out of the starting node to the total network flow flowing into the target node under a specific path.

[0132] Node output distribution matrix :

[0133] In the formula: It is the identity matrix; The system node flux matrix; This is the power distribution matrix of the system branch outflow.

[0134] Path output distribution matrix :

[0135] In the formula: Inject a distribution matrix into the system units; It is a row vector consisting entirely of 1s. To constitute The vector of active power of the generator.

[0136] (2) Relationship between wind and solar power output fluctuations and the response of balanced units According to the real-time power balance constraints of the power system, the random fluctuations in wind and solar power output and the resulting changes in system network losses must be borne by the system balancing units. Assuming the load power remains constant, the balance relationship is as follows:

[0137] Adjustment of output of the balancing unit for:

[0138] In the formula: , These represent the fluctuations in power output of wind farms and photovoltaic power plants, respectively. This represents the fluctuation in the system's active power loss; the negative sign indicates that the adjustment direction of the balancing unit is opposite to the fluctuation direction of the new energy source.

[0139] (3) Definition of the influence factors of the balancing unit on the carbon flow of the system The influence factor vector of the output change of the balancing unit on the carbon flow rate of nodes, branches, and network losses in the system is defined as follows: Balanced Units—Nodal Carbon Flow Influence Factor Vector :

[0140] In the formula: To balance the carbon emission intensity of the unit; To represent the balance node Indicator vector (the first) Behavior 1, the rest are 0). The carbon flow correlation vector for the balancing unit-node.

[0141] Balanced unit—branch carbon flow influence factor vector :

[0142] in, This is the carbon flow correlation vector between the balancing unit and the branch.

[0143] Balanced Units—Network Loss Carbon Flow Influence Factor Vector :

[0144] In the formula: This is the active power loss matrix for the branch. The carbon flow correlation vector between the balancing unit and the network loss.

[0145] (4) Expression for the influence of wind and solar power input on carbon flow distribution The changes in system node carbon flow rate, branch carbon flow rate, and network loss carbon flow rate caused by wind and solar power input are calculated as follows: Changes in nodal carbon flux due to wind turbine injected power Changes in carbon flow rate of branch circuits and the change in carbon flow rate of network loss Impact:

[0146]

[0147]

[0148] The change in node carbon flux due to photovoltaic power injection Changes in carbon flow rate of branch circuits Changes in carbon flow rate and network loss Impact:

[0149]

[0150]

[0151] Changes in node carbon flux due to combined wind and solar injection power Changes in carbon flow rate of branch circuits and the change in carbon flow rate of network loss Impact:

[0152]

[0153]

[0154] (5) Calculation and extraction of average impact factor To obtain the average impact factor under wind and solar uncertainties, progressive Latin hypercube sampling (PLHS) was used to analyze wind speed and light intensity. The second sampling yielded a series of wind and solar power output fluctuation scenarios. , ( =1,2,……, ); whereby, the fluctuation amount in the wind and solar power output fluctuation scenario is defined as the difference between the current sample power output value obtained by PLHS sampling and the mathematical expectation value of the variable (i.e. This reflects the degree of deviation of wind and solar power output from the statistical mean.

[0155] The average impact factor of wind power on nodes, branches, and network losses is calculated as follows:

[0156]

[0157]

[0158] In the formula: , , They are respectively The average impact factor of wind power on node, branch, and network loss carbon flow rate after round sampling.

[0159] The average impact factor of photovoltaics on nodes, branches, and grid losses is calculated as follows:

[0160]

[0161]

[0162] In the formula: , , They are respectively The average impact factor of photovoltaic on node, branch and network loss carbon flow rate after round sampling.

[0163] The average impact factor of wind-solar combined operation on nodes, branch lines, and network losses is calculated as follows:

[0164]

[0165]

[0166] In the formula: , , They are respectively The average impact factor of wind and solar power on the carbon flow rate of nodes, branches, and network losses after round sampling.

[0167] Example 2: To evaluate the superiority and effectiveness of the proposed PLHS sampling strategy and the collaborative optimization recombination of sample sequences, such as Figure 3 As shown, this invention utilizes the IEEE 14-node system for probabilistic optimal power flow calculation. The example uses 35,000 sets of wind speeds (approximately following a Weibull distribution) and solar irradiance (approximately following a Beta distribution) from a provincial power system to analyze the output variables of the probabilistic power flow. The main parameters for wind power and photovoltaic power are shown in the table below. Wind farm WF1 is connected to node 3 and photovoltaic power station PV1 is connected to node 2.

[0168] Table 1. Basic requirements and some equipment parameters for this invention

[0169] Table 2 Parameters of IEEE 14-Node System

[0170] Table 3. Some parameters of wind farms and photovoltaic farms in the power system.

[0171] To demonstrate the effectiveness and superiority of PLHS-based probabilistic power flow, node voltage, branch active power, and system network losses are analyzed as output variables. Using the results of 10,000 MCS iterations as a reference, 1,000 LHS and PLHS iterations are performed for comparative analysis. In this example, the PLHS sampling uses 10 slices, with each slice containing 100 samples.

[0172] In probabilistic power flow calculations based on PLHS, random variables encompass uncertainties in wind power output and solar power output. In practice, PLHS is first used to sample key parameters such as wind speed, solar irradiance, and load, successfully acquiring 1000 initial sample data sets. These initial sample data are then input into the IEEE 14-bus system to perform deterministic power flow calculations. The final voltage amplitude power flow calculation results are detailed in the table below.

[0173] Node 1 is designated as the slack node, while nodes 2, 3, 6, and 8 are configured as PV nodes, with their voltage amplitudes remaining constant; therefore, they are not listed in the table. Monte Carlo simulation (MCS) sampling results are used as a benchmark. Data analysis shows that the mean voltage of each node in the probabilistic power flow calculation output based on PLHS perfectly matches the MCS benchmark value, and the variance characteristics reflect the inherent fluctuation characteristics of the probabilistic analysis method. This comparison verifies the computational effectiveness of the PLHS method in probabilistic power flow analysis of power systems. Its statistical characteristics are highly consistent with the classic MCS algorithm, and the deviations are all controlled within the engineering allowable error threshold.

[0174] Table 4 Comparison of Calculation Results of Voltage Amplitude at IEEE 14-Node

[0175] PLHS effectiveness assessment in Example 2: Figure 4 and Figure 5 The probability density and cumulative distribution of active power in branches 12-13 are compared under three methods. Analysis shows that the probability density of active power in branches 12-13 obtained by the PLHS method is in high agreement with that obtained by the MCS method, while the results of the Latin hypercube sampling method (LHS) show some deviation. This is further supported by the quantitative indicators in Table 4. Figure 6 The cumulative probability distribution characteristics of the total network loss of the system clearly show that the probability power flow calculation method based on PLHS has significant advantages in both accuracy and reliability, fully verifying its engineering applicability.

[0176] Evaluation of the superiority of PLHS in Example 2: Taking the voltage amplitude at node 12 as the research object, the mean and variance were determined as benchmark parameters based on 10,000 probabilistic power flow calculations using the MCS. Further, the deviations of the mean and variance obtained from every 100 probabilistic power flow calculations relative to the benchmark values ​​were compared.

[0177] For the probabilistic power flow calculation problem, three sampling methods were used for analysis, with parameter settings in steps of 100. The expected value and variance relative error results of the voltage amplitude at node 12 are shown below. Figure 7 , Figure 8 In summary, ignoring computational errors and the inherent volatility of probabilistic power flow calculations, all three methods exhibit a trend of decreasing error with increasing sampling size. Among them, the PLHS method significantly outperforms the LHS method: under the same sampling size, the PLHS method shows smaller relative errors in mean and variance, thus verifying the superiority of the PLHS method in reducing estimation errors.

[0178] PLHS efficiency evaluation in Example 2: The efficiency of PLHS is evaluated by comparing the probabilistic power flow calculation time of the three methods above, as shown in Table 5.

[0179] The table below lists the time required for probabilistic power flow calculation using three methods. The comparison shows that the PLHS-based probabilistic power flow calculation method proposed in this paper has a single-sample computation time increase of only 0.28ms compared to LHS. This is because PLHS introduces an optimal slice sorting operation on top of LHS. Furthermore, under the same accuracy requirements, PLHS has a faster convergence speed than MCS, requiring less computation time and thus improving computational efficiency.

[0180] Comprehensive analysis shows that the PLHS-based probabilistic power flow calculation method exhibits significant advancements and superiority in both computational accuracy and efficiency. By optimizing sampling strategies and probabilistic modeling mechanisms, this method effectively improves the computational efficiency of probabilistic analysis of complex power systems while ensuring the reliability of results. It provides a more computationally efficient technical solution for large-scale uncertainty assessment of new energy sources and lays the foundation for uncertain carbon flow analysis.

[0181] Table 5. Efficiency Comparison of Three Probabilistic Power Flow Calculation Methods

[0182] Example 3: To verify the accuracy of the correlation model between wind-solar synergy and the total carbon flow rate injection of the system, as well as the carbon flow influencing factors under wind-solar uncertainty, this embodiment continues to use the IEEE 14-bus system as the analysis object. Conventional generators and loads are connected to the grid architecture in a deterministic manner. The focus is on studying the impact mechanism of injected power uncertainty caused by wind speed and solar intensity fluctuations on the system's carbon flow distribution. Based on research needs, six typical operating scenarios were constructed. Latin hypercube sampling (PLHS) was used to sample wind and solar power output 1000 times, generating a statistically representative set of uncertainty scenarios. On this basis, carbon flow distribution calculations were performed, systematically analyzing the impact of wind-solar power output fluctuations on carbon flow paths, node carbon potential, and branch carbon flow rates, providing theoretical support for the low-carbon operation of power systems with a high proportion of renewable energy.

[0183] The wind turbine capacity connected to the grid is 50 MW and 100 MW, and the photovoltaic capacity connected to the grid is also 50 MW and 100 MW. Whether the wind turbines are generating power or not, the carbon emission intensity of both wind and photovoltaic generators is 0. The carbon emission intensity of the generators is shown in the table below: Table 6 Carbon emission intensity of non-wind and solar generator units in the IEEE 14-bus system

[0184] If a wind turbine is connected, it replaces the current turbine. At this time, the carbon emissions are 0. For example, if the G3 turbine is completely replaced, its output becomes wind power and the carbon emission intensity is set to 0, other turbines need to adjust their output to balance the system. G1 is set as the balancing turbine and may reduce its output to accommodate the wind power injection. G2 / G4 / G5 adjust their output according to the system requirements. If a photovoltaic turbine is connected to the node, its output is increased on the basis of the original turbine. The increase in output is achieved by injecting a negative load to simulate the increase in output.

[0185] Table 7 Penetration Rate Settings for Different Wind and Solar Farm Scenarios

[0186] Carbon flow rate correlation analysis in Example 3: Figure 9 The results are from 1000 PLHS simulation wind speed samplings. In Scenario 1, since no wind and solar turbines are connected, the total carbon flow rate of the system remains constant at 224.3 tCO2 / h. The impact of wind speed changes on the total carbon flow rate of the system in Scenario 2 and Scenario 3 (e.g., ...) is discussed. Figure 10 As shown in the figure, when the wind farm is connected to system node 3, and the wind speed is between the cut-in wind speed and the rated wind speed, the total carbon flow rate of the system exhibits a cubic curve variation characteristic with wind speed fluctuations; when the wind speed exceeds the rated wind speed but does not reach the cut-out wind speed, the total carbon flow rate of the system drops to its trough. The above pattern is highly consistent with the derivation results of the total carbon flow rate of the system, verifying its effectiveness. Comparing the fitted curves of scenario 2 and scenario 3, it can be seen that the greater the wind power penetration level, the greater the fluctuation of the total carbon flow rate of the system due to the wind farm. The introduction of wind power injection introduces uncertainties into the system carbon flow analysis, and the increase in the rated power of the wind farm can significantly enhance the carbon reduction effect of the system, which provides a theoretical basis for setting the initial boundary conditions of the probabilistic carbon flow model.

[0187] Figure 11 This is the result of 1000 PLHS simulations of light intensity sampling. The correlation between light intensity and the total carbon flow rate of the system is as follows: Figure 12 As shown, the trend is similar to that of wind farms: comparing scenarios 4 and 5, as the light intensity gradually increases, the total carbon flow rate of the system gradually decreases, showing a near-linear correlation. When the light intensity reaches a specific threshold, the photovoltaic power output reaches its rated power, at which point the total carbon flow rate of the system drops to its lowest level. These results are consistent with the derived conclusions, verifying their effectiveness. Comparing the curves of scenarios 4 and 5, it can be seen that the higher the photovoltaic penetration level, the greater the decrease in the total carbon flow rate of the system, and the more significant the fluctuation.

[0188] Figure 13The carbon flow analysis results for Scenario 6 are presented. When the system is simultaneously connected to wind and solar power, operational volatility increases significantly. The total carbon flow rate is influenced by a dual coupling of wind speed and solar irradiance: when the wind speed is outside the cut-out wind speed range and the solar irradiance is at the minimum generating capacity of the solar panels, the total carbon flow rate is high; while when the wind speed is between the rated wind speed and the cut-out wind speed and the solar irradiance is relatively high, the total carbon flow rate decreases significantly. This phenomenon verifies the applicability of the model and is consistent with the actual operating patterns of the system.

[0189] Analysis of various scenarios in Example 3: Figure 14 The paper presents the branch carbon flow rate distribution of wind and solar power grid integration systems under six different scenarios. Each branch carbon flow rate is numbered according to the actual power flow direction and uniformly treated as a positive value. System power fluctuations are mainly borne by balancing node 1. When wind power replaces the original power source, the active power output of G1 is forced to increase. Since G1 has the highest carbon emission intensity, wind farm integration leads to an upward trend in branch carbon flow rate, with the impact of a 50MW wind farm integration being particularly significant on branches adjacent to G1. When solar power is integrated into the original power source in a superimposed output manner, comparing scenarios 1, 2, and 3, and scenarios 1, 4, and 5, the branch carbon flow rate shows a significant downward trend. Further comparing scenarios 1, 2, 5, and 6, the branch carbon flow rate generally decreases. This is mainly due to two factors: first, the increase in wind and solar power injection reduces the total system carbon flow rate; second, the increased output of wind farm G3 reduces the output demand of G1, thus leading to a decrease in branch carbon flow rate. In summary, the main influencing factor on the carbon flow rate of a branch is the active power output level of G1, and the carbon flow rate decreases more significantly in branches farther from the power source. For example... Figure 15 The figure shows the total carbon flow rate of the branches in each scenario. Specifically, the total carbon flow rate of the branches in scenario six is ​​524.78 tCO2 / h, which is 6.54% lower than that in scenario five (535.14 tCO2 / h) and 13.43% lower than that in scenario four (537.81 tCO2 / h).

[0190] Depend on Figure 16 Analysis shows that, consistent with the findings of the branch carbon flow rate study, the carbon potential of the connected nodes is significantly reduced when a 100MW wind turbine is connected to node 3 or a photovoltaic unit is connected to node 2. This indicates that the integration of new energy sources effectively reduces node carbon potential, and that adding new power output is more effective than replacing existing units. This conclusion is also validated by the comparison results between scenarios 2 and 3, showing that node carbon potential is significantly affected by wind power penetration. Adding wind and solar power directly to the existing system significantly reduces the carbon potential of each node because the added photovoltaic power output directly reduces the power demand of G1, thereby lowering the node carbon potential. Figure 17The figure shows the average carbon potential of nodes under each scenario. Specifically, the average carbon potential of node 6 in scenario 6 is 761.32 kgCO2 / (MW·h), which is 8.99% lower than that in scenario 1 (836.56 kgCO2 / (MW·h)), 2.09% lower than that in scenario 3 (777.63 kgCO2 / (MW·h)), and 6.77% lower than that in scenario 4 (816.63 kgCO2 / (MW·h)).

[0191] Influence factor analysis in Example 3: Table 8 Unit load contribution based on power flow tracing method

[0192] Table 9. Unit network loss contribution based on power flow tracing method

[0193] Based on the proportional allocation principle of the countercurrent power flow method, active power flow source analysis was performed on the system of scenario 6 to obtain the power contribution and carbon flow rate contribution of the three thermal power units to each load node (of which node 1, node 7 and node 8 are non-load nodes). Table 4 shows the contribution of the units to each load and the total network loss. Since units G2, G3 and G5 are zero-carbon units, their influence factors are all 0, so they are not given in the table.

[0194] Select any set of data from Table 8. For example, the contribution of unit G1 to load 6 in terms of carbon flow rate is 4.413 tCO2 / h, and the contribution of active power is 5.04 MW. The ratio of the two is 0.875 tCO2 / (MW·h), which is the carbon emission intensity of unit G1. This verifies the correctness of the lossy network calculation method based on the countercurrent power flow method proposed in this paper. Next, the balance factor is calculated based on the results of unit load and network loss contribution based on the power flow tracing method.

[0195] Next, we analyze the average impact factor of the system units on the nodes. Average influence factor on carbon flow rate of system branches and carbon flow rate of system branch network loss ,in, , , .

[0196] Table 10 Average Influence Factors of Unit Injected Power on System Node Carbon Flow Rate

[0197]

[0198] Table 11 Average Influence Factor of Unit Injected Power on Carbon Flow Rate of System Branches

[0199]

[0200] This invention, based on the countercurrent tidal method and the carbon flow influence factor theory, interprets the carbon flow influence factors of nodes and branches calculated in the invention. The core physical meaning of the influence factor is: the change in carbon flow rate of a system node or branch caused by a change in the unit output of the balancing unit. A negative value indicates that an increase in wind and solar power output (leading to a decrease in the output of the balancing unit) will reduce the carbon flow rate at that location.

[0201] Table 10 shows the average impact factor of carbon flow rate at each node when wind power (WT), solar power (PV), and wind-solar joint are connected. The new data shows a more significant gradient and spatial distribution characteristics, which are highly consistent with theoretical expectations. The joint impact factor of the balancing node is -1.0970. The balancing node is usually the connection point of high-carbon units (such as the G1 coal-fired unit). The increase in wind and solar power output leads to a decrease in the output of the balancing unit, and the direct result is a significant decrease in the carbon flow rate generated by the balancing node itself. Therefore, the impact factor of this node is negative and has a large absolute value. Selecting any value from the table, such as calculating the ratio of the carbon flow contribution to the power contribution of the G1 unit to any load, the result will necessarily be equal to the carbon emission intensity of G1 (0.875 tCO2 / MWh). This fundamental relationship still holds, proving that the entire analysis chain from power flow calculation to carbon flow tracking maintains inherent mathematical rigor and reliability.

[0202] In addition, the impact factor of photovoltaic (PV) power on node 2 (-0.8750) is much greater than that of wind power (WT) power on the same node (-0.4230). Conversely, the impact of wind power on node 3 (-0.8750) is much greater than that of PV power (-0.3890). This is because PV power plants are connected to node 2, while wind farms are connected to node 3. According to the principle of "nearest absorption" and the power flow distribution of the network, the output of each power source has the most direct and dramatic impact on the carbon flow rate of the node with the closest electrical distance, reflecting the shaping effect of network topology and power source layout on carbon flow distribution. Extending outward from the main power source points (nodes 1, 2, 3), such as to nodes 7, 8, 9, 10, etc., the absolute value of the impact factor generally shows a clear decreasing trend. For example, the joint impact factor gradually decreases from -1.0970 at node 1 and -1.2980 at node 2 to -0.6100 at node 14. This indicates that the impact of changes in the output of balancing units on carbon flow weakens with increasing electrical distance, which is consistent with the basic law of power flow distribution in power systems. Node 8 is isolated, with an influence factor of 0, indicating that there is no net active power flow crossing this node or it is a purely reactive node. Therefore, its carbon potential is not affected by changes in the output of the balancing unit.

[0203] Table 11 shows the average impact factor of wind and solar power output on the carbon flow rate of branches. The numerical changes more clearly reveal the "path" and "corridor" of carbon flow in the power grid. The impact factors of branches 15 and 17 (7-9, 9-14) are 0, consistent with the analysis of node 8, indicating that these branches may have very small power flows or be in a different direction from the main channel under the selected operating mode, hence their impact factor of 0. Branches connecting nodes with high impact factors also exhibit large absolute values ​​of their own impact factors. For example, branch 1 connecting nodes 1 and 2 has a joint impact factor as high as -0.8230, indicating that it is a key carbon flow corridor for transmitting high-carbon electricity from balancing node 1 to the load area of ​​node 2, and therefore is extremely sensitive to changes in wind and solar power output.

[0204] Table 12 Factors influencing the total grid loss carbon flow rate of wind and solar turbines

[0205]

[0206] Table 12 provides the impact factors of wind and solar power output on the carbon emissions corresponding to total active power loss from a system-wide perspective. All impact factors are negative, indicating that as wind and solar power output increases and balancing unit output decreases, the system's total active power loss and corresponding carbon emissions also decrease, quantifying the additional environmental benefits brought by wind and solar integration—reducing carbon emissions from grid losses.

[0207] In addition, wind and solar power differ in their influencing factors. Wind power has a slightly higher impact factor on grid losses than solar power, possibly because the output fluctuations of wind farms have a more significant effect on network power flow due to differences in grid connection nodes and their own capacity, thus having a greater impact on grid losses. The combined effect is approximately the sum of the two, indicating that wind and solar power have a certain cumulative effect in reducing carbon emissions from grid losses.

[0208] In summary, the significant negative impact factor of the balancing node, the extreme value effect of the renewable energy access point, the attenuation effect with electrical distance, the identification of key carbon flow corridors, and the quantification of system network loss carbon flow all align with and deepen the carbon emission flow theory established in this invention. The calculation results have passed multiple internal verifications, fully demonstrating the correctness, robustness, and practicality of the probabilistic carbon flow model and calculation method.

[0209] This disclosure can be a system, method, and / or computer program product. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of this disclosure.

[0210] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example—but not limited to—electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination of the foregoing. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0211] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0212] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.

[0213] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A wind-solar-carbon flow correlation analysis method based on PLHS sampling and sample optimization, characterized in that, Includes the following steps: Step 1: Construct a probability distribution model for wind and solar power output; construct a comprehensive analysis model for system carbon flow rate under the coupling uncertainty of wind and solar power output; Step 2: A hierarchical Latin hypercube sampling method is used to sample wind speed and light intensity to obtain a sample sequence; an adaptive sequence search algorithm is proposed, which uses a composite index based on spatial dispersion as the optimization objective to optimize the sample sequence. The optimized sample sequence is then input into the wind and solar power output probability distribution model to calculate the wind and solar power output sample set; the optimization objective balances the minimization of Euclidean distance between sample points and the uniformity of projection of each variable dimension through weighting coefficients. Step 3: Traverse the sample points in the wind and solar power output sample set, and calculate the system carbon flow rate and its probability distribution characteristics based on the system carbon flow rate comprehensive analysis model; Step 4: Based on the reverse current method and the path output distribution factor theory, for each sample point in the wind and solar power output sample set, calculate the changes in system node carbon flow rate, branch carbon flow rate, and network loss carbon flow rate caused by wind and solar power output injection; and perform a weighted average to obtain the average influence factor of wind and solar power output on system node carbon flow rate, branch carbon flow rate, and network loss carbon flow rate.

2. The wind-solar-carbon flow correlation analysis method based on PLHS sampling and sample optimization according to claim 1, characterized in that: The comprehensive analysis model of system carbon flow rate under the uncertainty of wind and solar power output coupling is used to quantify the mapping relationship between the wind and solar power output with random fluctuations and the total carbon flow rate of the system, specifically: in, The total carbon flow rate of the system. For the combined wind and solar power output vector, It is a unit vector; For the first s The carbon emission intensity of the node, and the first node s Each node is a node for the balanced generator unit access; is a constant, representing all deterministic factors other than wind and light fluctuations; For the first, excluding the balancing unit The generator injection power of each node, with this value being 0 for load nodes; For the first The load power of each node; , They are nodes j ,node m Load power, node j Indicates the wind farm access node, node m Indicates the grid connection node of a photovoltaic power station; This is due to network loss in the system. , The first k Carbon emission intensity and active power of each node connected to the unit; The combined wind and solar power output vector Specifically: ,in, For the output power of the wind farm, Wind speed; For the output power of photovoltaic power plants, Solar irradiance.

3. The wind-solar-carbon flow correlation analysis method based on PLHS sampling and sample optimization according to claim 2, characterized in that: The total carbon flow rate of the system Let be a random variable driven by combined wind and solar power output, with its expected value and variance as follows: in, Let be the mathematical expectation of the total carbon flow rate of the system; Let V be the variance of the total carbon flow rate of the system; , These are the mathematical expectations of the output power of the wind farm and the output power of the photovoltaic power station, respectively. , These are the variances of the output power of the wind farm and the output power of the photovoltaic power station, respectively. Covariance contributing to wind and solar power.

4. The wind-solar-carbon flow correlation analysis method based on PLHS sampling and sample optimization according to claim 1, characterized in that: In step 2, the objective function for maximizing spatial dispersion is as follows: in, It represents a sequence of sample subsets, that is, a permutation and combination; , They are respectively i The, the j One sample point; The Euclidean distance between sample points; D represents the total number of all sample points; D represents the total number of dimensions of the variables. The number of intervals divided for each dimension variable; For the first The dimension variable falls on the th The number of samples in each interval; , , which are weighting coefficients used to balance the two objectives of distance between points and projection uniformity.

5. The wind-solar-carbon flow correlation analysis method based on PLHS sampling and sample optimization according to claim 4, characterized in that: Step 2, the adaptive sequence search algorithm, specifically includes the following steps: Step 2.1: Randomly generate an initial sample sequence and calculate its objective function value; Step 2.2: Define the neighborhood operation as swapping the positions of any two sample subsets in the sample sequence, and then perform an iterative search, specifically: Step 2.2.1: After performing the neighborhood operation on the current sample sequence, a new sample sequence is generated. If the objective function value of the new sample sequence is better, it is directly accepted as the current optimal solution. Otherwise, the probability of acceptance is calculated according to the Metropolis criterion. P The probability is combined with a random number. In comparison, if If the objective function value temporarily deteriorates, then the new sample sequence that causes it to deteriorate is accepted as the current optimal solution to avoid getting trapped in a local optimum; if Then the original sample sequence is taken as the current optimal solution, where, satisfy ; Step 2.2.2: Using the sample sequence corresponding to the current optimal solution as a benchmark, perform a fine search only on the neighborhood sample sequences that can be generated through adjacent subset swapping and subset position sliding operations, until continuous m The objective function value was not improved in the second instance, where, m This is the threshold for the number of iterations; Step 2.2.3: If continuous N glo If no better solution is found in the next iteration, the perturbation amplitude is increased to enhance the global exploration capability; if the near... N loc If the objective function value continuously increases in each iteration, then the perturbation amplitude is reduced, and a finer search is performed; where... N glo The threshold for triggering global exploration, N loc This is the threshold for triggering a local fine-tuning search. Step 2.3: When the preset maximum number of iterations is reached... C limit Or the objective function value is in continuous L The iteration terminates when the relative improvement is less than a preset relative threshold; where the relative improvement is the difference between the objective function value of the current iteration and the objective function value of the previous iteration divided by the absolute value of the objective function value of the previous iteration.

6. The wind-solar-carbon flow correlation analysis method based on PLHS sampling and sample optimization according to claim 5, characterized in that: The acceptance probability is calculated based on the Metropolis criterion. P The calculation method is as follows: in, The objective function value for the current sample sequence; The objective function value of the new sample sequence after perturbation; For the first k The control parameters at the time of the next iteration, and as the number of iterations increases, According to the attenuation coefficient Gradually decrease, that is , satisfy ; In step 2.2.3, increasing the disturbance amplitude specifically includes increasing the temperature attenuation coefficient. The reduction of disturbance amplitude includes reducing the temperature attenuation coefficient. .

7. The wind-solar-carbon flow correlation analysis method based on PLHS sampling and sample optimization according to claim 2, characterized in that: In step 4, the changes in system node carbon flow rate, branch carbon flow rate, and network loss carbon flow rate caused by the wind and solar power output injection are calculated as follows: in, , These represent the changes in output power of wind farms and output power of photovoltaic power plants, respectively. , , These represent the changes in carbon flow rate at system nodes, carbon flow rate at branch lines, and carbon flow rate at network losses, respectively. To balance the influence factor vector of node carbon flow on the unit, To balance the influence factor vector of the unit on the branch carbon flow, The vector of factors affecting the carbon flow of the balancing unit on network losses.

8. The wind-solar-carbon flow correlation analysis method based on PLHS sampling and sample optimization according to claim 7, characterized in that: The influence factor vector of the balancing unit on the node carbon flow is determined in the following manner: in, To balance the carbon emission intensity of the unit; To represent the balance node s Indicator vector, Output the distribution matrix for the carbon flow path; The influence factor vector of the balancing unit on the branch carbon flow is determined in the following manner: in, The node output distribution matrix for carbon flow; The influence factor vector of the balancing unit on network loss carbon flow is determined in the following manner: in, It is a row vector consisting entirely of 1s. This is the active power loss matrix for the branch. This is the system node flux matrix.

9. The wind-solar-carbon flow correlation analysis method based on PLHS sampling and sample optimization according to claim 8, characterized in that: In step 4, the average influence factor of the wind and solar power output on the carbon flow rate of system nodes, branch carbon flow rate, and network loss carbon flow rate is determined as follows: The average impact factors of wind power output on system node carbon flow rate, branch carbon flow rate, and network loss carbon flow rate are as follows: in, , , The hierarchical Latin hypercube sampling method was used for the following purposes: The average impact factor of wind power output on system node carbon flow rate, branch carbon flow rate and network loss carbon flow rate after round sampling; K For the total number of sampling rounds; , , The first k The vectors of the balancing unit's influence on node carbon flow, the vectors of the balancing unit's influence on branch carbon flow, and the vectors of the balancing unit's influence on network loss carbon flow are corresponding to the subsamples. , The first k The changes in the output power of wind farms and photovoltaic power plants corresponding to the sub-samples; The average impact factors of photovoltaic output on the carbon flow rate of system nodes, branch carbon flow rate, and network loss carbon flow rate are as follows: in, , , They are respectively The average impact factor of photovoltaic output on system node carbon flow rate, branch carbon flow rate and network loss carbon flow rate after round sampling.

10. A wind-solar carbon flow correlation analysis system based on PLHS sampling and sample optimization according to any one of claims 1-9, comprising a system carbon flow rate comprehensive analysis model construction module, a wind and solar output sample set acquisition module, a system carbon flow rate calculation module, and a wind and solar injection carbon flow rate impact analysis module, characterized in that: The module for constructing a comprehensive analysis model of system carbon flow rate is used to build a probability distribution model of wind and solar power output and a comprehensive analysis model of system carbon flow rate under the coupling uncertainty of wind and solar power output. The wind and solar power output sample set acquisition module uses a hierarchical Latin hypercube sampling method to sample wind speed and light intensity to obtain a sample sequence. An adaptive sequence search algorithm is proposed, which uses a composite index based on spatial dispersion as the optimization objective to optimize the sample sequence. The optimized sample sequence is then input into the wind and solar power output probability distribution model to calculate the wind and solar power output sample set. The optimization objective balances the minimization of Euclidean distance between sample points and the uniformity of the projection of each variable dimension through weighting coefficients. The system carbon flow rate calculation module traverses the sample points in the wind and solar power output sample set and calculates the system carbon flow rate based on the system carbon flow rate comprehensive analysis model. The wind and solar power injection carbon flow rate impact analysis module, based on the reverse tidal current method and path output distribution factor theory, calculates the changes in system node carbon flow rate, branch carbon flow rate, and network loss carbon flow rate caused by wind and solar power injection for each sample point in the wind and solar power output sample set; and performs a weighted average to obtain the average impact factor of wind and solar power on system node carbon flow rate, branch carbon flow rate, and network loss carbon flow rate.

Images