Environmental-economic distributed robust dispatch method considering the correlation of renewable energy sources

Abstract

The present invention relates to the technical field of power regulation, and in particular to an environmental economic distribution robust scheduling method considering the relevance of renewable energy, including: establishing an R-copula joint distribution function of multiple renewable energy power stations; performing interval prediction on the power of each renewable energy power station to obtain a power interval with multiple confidence levels; extracting a preset number of renewable energy power data for clustering according to the power interval with multiple confidence levels and the R-copula joint distribution function, generating multiple typical renewable energy output scenarios, and inputting them into a target scheduling center for environmental economic distribution robust optimization, while controlling carbon emissions and atmospheric pollutant emissions, so as to obtain a start-stop plan for thermal power units and an optimal value for resource allocation, and based on this, formulating a scheduling plan that satisfies the power balance of the power grid. Thus, the problems of rough pollutant control methods and insufficient consideration of the relevance and uncertainty of renewable energy in traditional environmental economic scheduling are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of electric power regulation and control, and in particular to an environmentally economically robust scheduling method considering the relevance of renewable energy. Background Art

[0002] Global warming and air pollution are urgent problems that the world needs to solve. Traditional power system economic dispatch often focuses on maximizing economic benefits and pays relatively little attention to environmental impact. The rapid development of renewable energy such as wind and light can significantly change the power structure dominated by coal-fired power and control environmental pollution from the source. However, multiple wind farms and photovoltaic power stations in the same region are affected by meteorological factors, and their power generation characteristics have certain correlations and uncertainties, which brings huge challenges to environmental economic dispatch. Therefore, it is of great significance to vividly describe the power generation characteristics between multiple renewable energy power stations and reasonably apply them in the environmental economic dispatch of power systems.

[0003] Copula functions are often used to describe the correlation between the outputs of renewable energy power stations. They are mainly divided into two categories: the elliptical distribution family and the Archimedean distribution family. The copula function of the elliptical distribution family can describe specific symmetric structures, but cannot capture the correlation structure caused by meteorological factors and other factors well. The copula function of the Archimedean distribution family can express more complex correlations, but is only applicable to two-dimensional variables. Therefore, technicians have proposed a vine structure copula function that decomposes the copula of multiple variables into a cascade model with a tree structure, which can better establish the correlation between multidimensional variables. However, the renewable energy output scenario directly generated by the vine structure copula function lacks combination with the probability forecast of the day-ahead stage, and it is easy to generate output scenarios that deviate from the actual situation.

[0004] The high proportion of renewable energy output in the power grid has a strong uncertainty, and its impact on the dispatch plan cannot be ignored. Optimization methods are usually used to deal with this uncertainty, including robust optimization, stochastic optimization and distributionally robust optimization methods. Distributionally robust optimization methods combine the characteristics of the first two and can balance the economy and robustness of the dispatch plan, so they are widely used. However, traditional distributionally robust optimization methods use moment information to describe probability distribution, which has limited information and requires semidefinite programming for solution, which is computationally complex.

[0005] Therefore, there is an urgent need for a dispatching method that can fully utilize data and solve the model quickly and accurately to ensure the economy, safety and environmental protection of the power grid environmental economic dispatching model. Summary of the invention

[0006] The present invention provides an environmental economic distribution robust scheduling method considering the relevance of renewable energy, so as to solve the problems of rough pollutant control mode and insufficient consideration of the relevance and uncertainty of renewable energy in traditional environmental economic scheduling.

[0007] The first aspect of the present invention provides an environmentally economically distributed robust scheduling method that takes into account the correlation of renewable energy sources, comprising the following steps: establishing an R-copula joint distribution function for a plurality of renewable energy power stations; performing interval prediction on the power of each renewable energy power station to obtain a power interval with multiple confidence levels; extracting a preset number of renewable energy power data according to the power interval with multiple confidence levels and the R-copula joint distribution function, clustering the preset number of renewable energy power data, and generating a plurality of typical renewable energy output scenarios; inputting the plurality of typical renewable energy output scenarios into a target scheduling center for environmentally economically distributed robust optimization to obtain an environmentally economically distributed robust optimization model; using the environmentally economically distributed robust optimization model to control carbon emissions and emissions of atmospheric pollutants to obtain a start-stop plan and optimal resource allocation value for thermal power units, and formulating a scheduling plan that satisfies the power balance of the power grid based on the start-stop plan and the optimal resource allocation value.

[0008] Optionally, the step of establishing an R copula joint distribution function of multiple renewable energy power stations includes:

[0009] Acquire historical power data of multiple renewable energy power stations, and collect historical power data of the multiple wind and solar power stations to obtain edge condition distribution;

[0010] According to the marginal condition distribution, an R-copula joint distribution function is constructed to describe the spatial correlation of outputs among multiple renewable energy power stations in the same region.

[0011] Optionally, performing interval prediction on the power of each renewable energy power station to obtain interval prediction results with multiple confidence levels includes:

[0012] Based on the LSTM algorithm, historical prediction results are predicted according to the historical power data of the multiple wind and solar power stations;

[0013] Calculating a prediction error based on the historical prediction results and the historical power data of the plurality of wind and solar power stations;

[0014] The ABKDE method is used to statistically fit the prediction errors to generate the interval prediction results with multiple confidence levels.

[0015] Optionally, the adopting the ABKDE method to statistically fit the prediction error to generate the interval prediction result with multiple confidence levels includes:

[0016] The ABKDE method is used to fit the prediction errors in different power intervals, and the prediction error probability density curves in different power intervals are obtained respectively;

[0017] The prediction error probability density curves in the different power intervals determine the multiple confidence interval prediction results.

[0018] Optionally, extracting a preset number of renewable energy power data according to the multi-confidence power interval and the R-copula joint distribution function, clustering the preset number of renewable energy power data, and generating a typical output scenario includes:

[0019] Generating random power samples uniformly in the R-copula joint distribution function;

[0020] Extracting power samples of the power interval that meets the multiple confidence levels from the random number power samples by an inverse transformation sampling method, and resampling power samples of the power interval that do not meet the multiple confidence levels until a preset number of samples is reached, thereby obtaining the preset number of renewable energy power data;

[0021] The preset number of renewable energy power data are clustered using a fuzzy clustering algorithm to generate the typical renewable energy output scenario.

[0022] Optionally, the use of the environmental economic distribution robust optimization model to control carbon emissions and emissions of atmospheric pollutants to obtain a start-stop plan and optimal resource allocation for thermal power units includes:

[0023] The environmental-economic distributed robust optimization model is used to control carbon emissions and emissions of atmospheric pollutants, and the column and constraint generation algorithm is used to perform distributed robust optimization on the multiple typical renewable energy output scenarios until the column and constraint generation algorithm converges to obtain the start and stop plan of the thermal power unit and the optimal value of resource allocation.

[0024] The second aspect of the present invention provides an environmentally economically distributed robust scheduling device that takes into account the correlation of renewable energy sources, including: a construction module for establishing an R-copula joint distribution function of multiple renewable energy power stations; a prediction module for performing interval prediction on the power of each renewable energy power station to obtain a power interval with multiple confidence levels; a clustering module for extracting a preset number of renewable energy power data according to the power interval with multiple confidence levels and the R-copula joint distribution function, clustering the preset number of renewable energy power data, and generating multiple typical renewable energy output scenarios; a robust optimization module for inputting the multiple typical renewable energy output scenarios into a target scheduling center for environmentally economically distributed robust optimization to obtain an environmentally economically distributed robust optimization model; a formulation module for using the environmentally economically distributed robust optimization model to control carbon emissions and emissions of atmospheric pollutants to obtain a start-stop plan and optimal resource allocation value for thermal power units, and formulate a scheduling plan that satisfies the power balance of the power grid based on the start-stop plan and the optimal resource allocation value.

[0025] A third aspect of the present invention provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement an environmentally economically distributed robust scheduling method considering the relevance of renewable energy as described in the above embodiments.

[0026] A fourth aspect of the present invention provides a computer program product, which, when executed by a processor, implements the above-mentioned environmentally economically robust scheduling method considering the relevance of renewable energy.

[0027] A fifth aspect of the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-mentioned environmentally-economically distributed robust scheduling method considering the relevance of renewable energy.

[0028] The environmentally economic distributed robust scheduling method considering the correlation of renewable energy proposed in the embodiment of the present invention solves the problems faced in the environmental economic scheduling of traditional power grids. On the one hand, for the control of carbon emissions and atmospheric pollutant emissions, total control and ground concentration control considering the gas diffusion model are adopted respectively; on the other hand, for the challenges of correlation and uncertainty brought by the renewable energy access system, by establishing a correlation model, the method can more accurately predict the output characteristics of renewable energy, thereby providing more reliable data support. The distributed robust optimization method is adopted, which fully considers the uncertainty of renewable energy output, improves the economy and reliability of the scheduling plan, and better meets the needs of environmental economic scheduling of power systems.

[0029] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0031] Figure 1 A flowchart of an environmental economic distribution robust scheduling method considering the relevance of renewable energy provided by an embodiment of the present invention;

[0032] Figure 2 A schematic diagram of the execution of the environmental economic distribution robust scheduling considering the relevance of renewable energy provided by an embodiment of the present invention;

[0033] Figure 3 A schematic diagram of the R-vine structure provided by an embodiment of the present invention;

[0034] Figure 4 LSTM neural network structure diagram provided by the embodiment of the present invention;

[0035] Figure 5 A schematic diagram of the overall execution of the environmental economic distribution robust scheduling considering the relevance of renewable energy provided by an embodiment of the present invention;

[0036] Figure 6 A block diagram of an environmental economic distribution robust scheduling device considering the relevance of renewable energy provided by an embodiment of the present invention;

[0037] Figure 7 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0038] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0039] The following describes an environmental economic distribution robust scheduling method considering the relevance of renewable energy according to an embodiment of the present invention with reference to the accompanying drawings.

[0040] Figure 1 A flow chart of an environmentally economically distributed robust scheduling method considering the relevance of renewable energy provided by an embodiment of the present invention.

[0041] like Figure 1As shown, the environmental economic distribution robust scheduling method considering the relevance of renewable energy includes the following steps:

[0042] In step S101, an R-copula joint distribution function of multiple renewable energy power stations is established.

[0043] In some embodiments, establishing an R copula joint distribution function of multiple renewable energy power plants includes:

[0044] Obtain historical power data of multiple renewable energy power stations, and collect historical power data of multiple wind and solar power stations to obtain edge condition distribution;

[0045] According to the marginal condition distribution, the R-copula joint distribution function is constructed to describe the spatial correlation of the outputs of multiple renewable energy power stations in the same region.

[0046] Specifically, Figure 2 As shown in the figure, the edge conditional distribution is obtained by obtaining the historical power data of multiple wind and solar power stations and performing statistics on them; the R-vine structure contains a series of trees, and in each tree, the Pearson correlation coefficient is used to obtain the correlation coefficient between the random variables of each node of the tree to ensure that the sum of the absolute values ​​of the Pearson correlation coefficient of each tree is the largest; then, according to the connection results of each node in each tree, the copula function with the best fitting effect between nodes is selected; finally, the node variables of each tree are connected through the vine structure to construct the R-vine copula joint distribution function, that is, to construct a spatial correlation model of multiple renewable energy power stations to describe the spatial correlation of the output between multiple renewable energy power stations in the same region. It should be noted that similar methods to the R-vine copula function include the C-vine and D-vine copula functions, but the C-vine and D-vine copula functions are fixed vine structures. In contrast, the R-vine structure is more flexible and can determine the vine structure according to the actual correlation between all variables. Figure 3 Schematic diagram of the R-vine joint distribution of five-dimensional random variables, whose connection relationship is not unique.

[0047] In the actual implementation process, there are N renewable energy stations, whose power variables can be expressed as {x1, x2, …, x N}Statistically, the marginal distribution function is {f1(x1), f2(x2), …, F N (x N )}, the R-vine structure can be used to reduce the dimensionality of high-dimensional random variables and establish the joint distribution function of high-dimensional random variables, that is, the spatial correlation model of multiple renewable energy power stations, as shown in formula (1).

[0048] H(x1, x2, ..., x N)=C(F1(x1), F2(x2),…,F N (x N )) (1)

[0049] Where H(·) is the joint distribution function of multidimensional variables, and C(·) is the copula distribution function.

[0050] The specific steps for establishing the correlation model are as follows:

[0051] (1) Determine the first tree of the R-vine copula function: For every two node data of the first tree, the Pearson correlation coefficient is used to obtain the correlation coefficient, as shown in formula (2).

[0052]

[0053] In the formula, cov(·) represents the covariance matrix, var(·) represents the variance of the random variable, and E(·) represents the mathematical expectation of the random variable. Calculate and sum the correlation coefficients of every two nodes to ensure that the sum of their absolute values ​​is the largest, so as to determine the node connection mode of the tree.

[0054] (2) Select the copula function with the best fitting effect: According to the connection results of the nodes in the previous step, the Akaike Information Criterion (AIC) is used to evaluate the goodness of the copula fitting of the connection nodes. Its expression can be expressed as:

[0055] AIC=2k-2ln(L) (3)

[0056] Where k is the number of parameters in the copula function, and L is the value of the maximum likelihood function of the copula function. The smaller the AIC, the better the model fit, and the optimal copula function for each group of nodes can be selected.

[0057] (3) Determine the vine structure of the next tree: According to the edge conditional distribution of formula (4), the conditional distribution value obtained based on the structure of the previous tree is used as the new node, and steps (1) and (2) are repeated until a complete R vine structure is constructed.

[0058]

[0059] In the formula, x and v are two random vectors, v i is the i-th variable in vector v, v -i Represents the vector v after removing the i-th variable.

[0060] In step S102, an interval prediction is performed on the power of each renewable energy power station to obtain power intervals with multiple confidence levels.

[0061] In some embodiments, the power of each renewable energy power station is predicted by interval prediction to obtain interval prediction results with multiple confidence levels, including:

[0062] Based on the LSTM algorithm, historical prediction results are predicted based on the historical power data of multiple wind and solar power stations;

[0063] Calculate the prediction error based on historical prediction results and historical power data of multiple wind and solar stations;

[0064] The ABKDE method is used to statistically fit the forecast error to generate interval forecast results with multiple confidence levels. Specifically,

[0065] The ABKDE method is used to fit the prediction errors in different power intervals, and the prediction error probability density curves in different power intervals are obtained respectively;

[0066] The prediction error probability density curves in different power intervals determine the interval prediction results with multiple confidence levels.

[0067] That is to say, Figure 2 As shown, based on the historical power data of multiple wind and solar power stations, the LSTM-ABKDE interval prediction technology is used to obtain the power intervals of multiple confidence levels of renewable energy stations in the day-ahead stage. It should be noted that other neural networks or algorithms can also be used to predict historical prediction results.

[0068] In the actual implementation process, Figure 4 As shown in the figure, firstly, an interval prediction model based on long short-term memory neural network (LSTM) is designed, which can be divided into two parts. The first part uses LSTM neural network to predict the power of each renewable energy site; the second part uses adaptive kernel density estimation (ABKDE) method to fit the distribution of prediction error and obtain the upper and lower limits of the interval under multiple confidence levels.

[0069] There are three gates inside the neuron, namely the forget gate, input gate and output gate. t Indicates the input at the current moment, h t and h t-1 Represent the states at time t and time t-1 respectively, C t and C t-1 They represent the states of the storage unit at time t and t-1 respectively. The calculation formulas of the internal state and output are shown in equations (5)-(10).

[0070] f t =σ(U f x t +W f h t-1 +b f ) (5)

[0071] i t =σ(U i x t +W i h t-1 +b i ) (6)

[0072] g t =φ(U g x t +W g h t-1 +b g ) (7)

[0073] o t =σ(U o x t +W o h t-1 +b o ) (8)

[0074]

[0075] In the formula, f t 、i t , g t , o t , C t and h t are the forget gate, input gate, input node, output gate, state unit and intermediate output state, respectively. f , U i , U g and U o are the corresponding gates and x t The matrix weights to be multiplied, W f , W i , W g and W o are the corresponding gates and h t-1 The matrix weights to be multiplied, b f , b i , b g and b o are the bias values ​​of the corresponding gates, Indicates bitwise multiplication of elements in the vector, σ indicates the change of sigmoid function, and φ indicates the change of tanh function. By setting the parameters of LSTM algorithm, taking historical power generation data as input, training and predicting through the network, the point prediction results of the day-ahead stage can be obtained.

[0076] The historical prediction results are subtracted from the actual power data to obtain a series of prediction errors. The ABKDE method is used to statistically calculate the prediction errors to obtain the probability density distribution, which is expressed as:

[0077]

[0078] Where k(x) is the window width h i is the nonparametric kernel density estimation function, λ is the bandwidth parameter, n is the number of renewable energy power point prediction error samples, α is the sensitivity factor, usually taken as 0.5. K(·) is the kernel function, and the Gaussian kernel function can be selected here, as shown in formula (13).

[0079]

[0080] The error distribution belonging to different power intervals is fitted by the ABKDE method, and the prediction error probability density curves of different power intervals are obtained respectively. The quantiles [0.05, 0.95], [0.10, 0.90] and [0.15, 0.85] of the error cumulative probability density curve are selected as the confidence intervals corresponding to confidence levels of 90%, 80% and 70%, respectively, to obtain the interval prediction results with multiple confidence levels.

[0081] In step S103, a preset number of renewable energy power data are extracted according to the power intervals with multiple confidence levels and the R-copula joint distribution function, and the preset number of renewable energy power data are clustered to generate a plurality of typical renewable energy output scenarios.

[0082] In some embodiments, a preset amount of renewable energy power data is extracted according to the power intervals of multiple confidence levels and the R-copula joint distribution function, and the preset amount of renewable energy power data is clustered to generate a typical output scenario, including:

[0083] Generate random power samples uniformly in the Rvine copula joint distribution function;

[0084] Power samples of power intervals satisfying multiple confidence levels are extracted from random number power samples by an inverse transformation sampling method, and power samples of power intervals not satisfying multiple confidence levels are resampled until a preset number of samples is reached, thereby obtaining a preset number of renewable energy power data;

[0085] A fuzzy clustering algorithm is used to cluster a preset number of renewable energy power data to generate typical renewable energy output scenarios.

[0086] Specifically, Figure 2 As shown, based on the R-vine copula model of step S101, the inverse transform sampling method is used for sampling, and the power samples that do not meet the confidence interval are resampled until the sampling number is reached; the fuzzy C-means clustering algorithm is used to cluster the massive power samples to obtain a typical renewable energy output scenario.

[0087] In the actual implementation process, for the power data at a certain moment, the power marginal distribution of the upper and lower bandwidths is predicted according to the interval, and the marginal distribution function value corresponding to the upper and lower bandwidths is obtained by using the interpolation algorithm. Subsequently, random numbers {a1, a2, ..., a n}Through the inverse transform sampling method, samples x1, x2, …, x n It can be expressed as:

[0088]

[0089] If the generated random number does not meet the marginal distribution function value range of the upper and lower power bandwidths, resampling is performed. Then the above process is repeated to generate renewable energy power data for the previous 24 hours.

[0090] For the massive output scenarios generated, the fuzzy C-means clustering algorithm is used to extract typical scenarios and the clustering loss function is defined as:

[0091]

[0092] In the formula, U is the membership matrix, V is the cluster center, N is the number of samples, C is the number of cluster centers, and u ij is the membership degree of a sample to the cluster center, m is the fuzzy index, x i is the i-th sample, c j is the jth cluster center. The clustering strategy is to iteratively calculate U and V to minimize J(U,V). The steps are as follows:

[0093] (1) Initialize the number of clusters C and the membership matrix U (0) . Let l = 1, which is the first iteration.

[0094] (2) Calculate V for the first iteration (l) .

[0095]

[0096] (3) Update the membership matrix U (l) , calculate the clustering loss function J (l) .

[0097]

[0098] (4) When or |J (l) -J (l-1) |<ε J Stop the iteration when ε u , ε J are the error thresholds of membership and loss functions respectively.

[0099] The final cluster center matrix V can represent the typical output scenarios of multiple renewable energy sources.

[0100] In step S104, a plurality of typical renewable energy output scenarios are input into the target dispatching center for environmental economic distribution robust optimization to obtain an environmental economic distribution robust optimization model;

[0101] In step S105, the environmental economic analysis robust optimization model is used to control carbon emissions and atmospheric pollutant emissions to obtain the start-stop plan and optimal resource allocation of the thermal power units, and a scheduling plan that satisfies the power balance of the power grid and is environmentally friendly is formulated based on the start-stop plan and the optimal resource allocation.

[0102] In some embodiments, the environmental economic distribution robust optimization model is used to control carbon emissions and atmospheric pollutant emissions to obtain the start-up and shutdown plan of the thermal power unit and the optimal value of resource allocation, including:

[0103] The environmental-economic distributed robust optimization model is used to control carbon emissions and atmospheric pollutant emissions, and the column and constraint generation algorithm is used to perform distributed robust optimization on multiple typical renewable energy output scenarios until the column and constraint generation algorithm converges to obtain the start and stop plans and optimal resource allocation of thermal power units.

[0104] Specifically, Figure 2 As shown, step S104 can be understood as constructing a data-driven environmental economic distribution robust optimization dispatching model for the power system. Environmental economic dispatch is to reduce carbon emissions and air pollution under the premise of ensuring the safe operation of the power grid. For carbon emissions, it can be controlled from the perspective of the total amount of carbon emissions generated by thermal power; for atmospheric pollutants, it should be controlled from the perspective of targeted pollutant ground concentration and diffusion. The optimization dispatching model is divided into two stages for modeling. In the first stage, the deterministic renewable energy output scenario is used as input to construct an objective function to minimize the start-up and shutdown costs, energy storage costs, carbon emission costs and system operation costs of the thermal power unit. At the same time, the start-up and shutdown constraints, climbing constraints, system power balance constraints, line safety constraints and atmospheric pollution environment constraints of the unit are considered, and the start-up and shutdown states of the unit are solved. In the second stage, the unit state obtained in the first stage is used as the input, and the 1-norm and ∞-norm are constructed based on the historical output data of renewable energy to constrain the probability of renewable energy output scenarios. Under the worst probability distribution, the optimal value of the pre-output and resource allocation of the thermal power unit is solved, and a dispatching plan that meets the power balance of the power grid and is environmentally friendly is formulated.

[0105] In the actual implementation process, the distributed robust optimization method is used to divide the power system environmental economic dispatch model into two stages for solution. In the first stage, the deterministic renewable energy output scenario is used as input, and the objective function can be divided into two parts. The first part is the start-up and shutdown cost of the thermal power unit, and the second part is the system operation cost, energy storage cost and carbon emission penalty cost, as shown in equations (19)-(21).

[0106] minF=minF1+minF2(19)

[0107]

[0108] In the formula, T is the dispatch period, G is the number of units, and F is on 、F off are the startup and shutdown costs of unit g at time t, u g,t is the start-up flag of unit g at time t. If it is 1, it means that unit g is started at time t; v g,t is the shutdown flag of unit g at time t. If it is 1, it means that unit g is shut down at time t; a g , b g and c g are the coal consumption coefficient matrices of thermal power units, P g,t is the power of unit g at time t, Q is the number of energy storage stations, are the charge and discharge amounts of the energy storage unit at time t, a c , b c and c c are the coefficient matrices of carbon emissions, ρ c is the penalty coefficient of carbon emission cost.

[0109] The constraint equations are shown in equations (22)-(39), where equation (22) is the unit start-stop constraint, equation (23) is the minimum start-stop time constraint, equation (24) is the output constraint, equation (25) is the ramp constraint, equation (26) is the power balance constraint, equation (27) is the line safety constraint, equations (28)-(33) are energy storage related constraints, equations (34)-(38) are the polluted gas Gaussian puff diffusion related models, and equation (39) is the polluted gas environment constraint. The specific expressions are as follows:

[0110]

[0111]

[0112] I c,t +I d,t ≤1 (30)

[0113] SOC s,min ≤SOC s,t ≤SOCs,max (31)

[0114]

[0115] SOC s,T =SOC s,0 (33)

[0116]

[0117] In the formula, x g,t is the operating status of thermal power unit g at time t, x g,t =1 means the unit is running; are the minimum start and stop time of unit g respectively; are the minimum and maximum output power of unit g respectively; R Ui , R Di are the climbing rate and sliding rate of the unit respectively; W and S are the number of wind farms and photovoltaic power stations respectively, w,t , P s,t are the wind and light powers at time t respectively; N is the total number of nodes, L n,t is the load power of node n at time t; S f,min , S f,max are the minimum and maximum transmission capacity matrices of the line, G l-g is the transfer distribution factor from bus g to line l, G l-w , G l-s , G l-q and G l-n Similarly; I c,t ,I d,t are symbolic variables for energy storage charging and discharging, P c,max , P d,max is the maximum energy storage for a single charge and discharge, E s,t is the energy storage charge at time t, SOC s,min , SOC s,max is the minimum and maximum value of charge, η c , η d is the energy storage charging and discharging efficiency, C GLPC (·) is the average pollutant concentration at the monitoring point, τ, t s are the emission time of polluted gas puff and the monitoring time respectively, J is the number of monitoring points, is the residence time of the polluted gas puff, G i (·) is the Gaussian diffusion function of the pollutant emission puff, x, y and z are the coordinates of the monitoring point, x o (τ,t s ),y o (τ,t s ) and z o(τ,t s ) is the smoke puff coordinate, σ x (τ,t s ),σ y (τ,t s ) and σ z (τ,t s ) is the diffusion parameter, x s ,y s and z s are the coordinates of the pollution source, u(t), v(t) and w(t) are the average wind speeds in three directions, Δt is the interval between two monitorings, α(t), β(t), λ(t) and γ(t) are the calculation coefficients of the diffusion parameters, which are determined by the atmospheric stability. It is the pollution concentration of the environmental boundary at the monitoring moment.

[0118] In the first stage model, the deterministic P w,t , P s,t As a constraint, the start-stop plan of the unit can be directly solved using the commercial solver Gurobi. In the second stage model, the uncertainty of renewable energy output needs to be considered. The above optimization scheduling model can be summarized as:

[0119]

[0120] In the formula, x is the decision variable in C1, that is, x g,t ;p r is the probability of each renewable energy output scenario, R is the number of discrete scenarios, Ξ is p r The uncertain set that satisfies ξ r is the renewable energy power value of the rth discrete scenario after scenario screening; y r is the decision variable in the rth discrete scenario C2, including P g,t , and Y(x,ξ r ) is y r The feasible domain that satisfies the constraints, c T ,d T are the cost coefficient matrices in C1 and C2 respectively; A, B, C, D, E, F and G are the coefficient matrices in the constraints, and e, h and g are constant column vectors. The uncertainty set Ξ can be specifically expressed as formula (42).

[0121]

[0122] In the formula, θ1, θ ∞ is the probability deviation limit corresponding to the 1-norm and ∞-norm, They are p r Relative to scene p r,0Positive and negative offsets of probability distributions, are 0-1 flag variables that generate positive and negative offsets, respectively. r,0 is the initial probability value of the rth scene.

[0123] To solve the above distributed robust optimization scheduling model, the column and constraint generation algorithm (C&CG) can be used to divide the model into main and sub-problems for solution. The main problem can be expressed as:

[0124] minc T x k +η k (43)

[0125]

[0126] Where superscript k is the number of iterations. The subproblem can be expressed as follows:

[0127]

[0128] Each discrete scenario of the sub-problem is independent of each other and can be solved directly. The specific solution steps using C&CG are as follows.

[0129] (1) The probability distribution obtained in fuzzy C-means clustering is used as the worst case. Initialize the lower bound LB = -∞, UB = +∞ and the number of iterations k = 1.

[0130] (2) Solve the main problem equation (43) to obtain x k , and update the lower bound LB = c T x k +η k .

[0131] (3) x k Substitute into the subproblem equations (45)-(46) and find the optimal solution to the subproblem and the worst case probability And update the upper bound

[0132] (4) Given a convergence threshold ε, if |UB-LB|<ε, the iteration stops and returns the optimal solution x k , and the worst probability of each scenario Otherwise, Add to the following constraints:

[0133]

[0134] Let the number of iterations k = k + 1, and then return to step (2) until the algorithm converges, and finally obtain the start-stop plan and scheduling arrangement of the unit.

[0135] like Figure 5 As shown, step S101 provides a correlation model to step S102, and steps S102 and S103 generate typical output scenarios according to the correlation model and the interval prediction results. Finally, step S103 provides data-driven support for step S104, and the distributed robust optimization method is used to solve the final scheduling arrangement.

[0136] In summary, the environmental economic distribution robust scheduling method considering the relevance of renewable energy proposed in the embodiment of the present invention has the following beneficial effects:

[0137] (1) The LSTM-ABKDE algorithm can be used to characterize the fluctuation range corresponding to the point prediction result under a predetermined confidence level. Compared with the traditional KDE method, the adaptive kernel density estimation (ABKDE) can use different bandwidths to complete the kernel density estimation for the power data in the window, thereby improving the accuracy and reliability of the estimation.

[0138] (2) Combining the R-vine copula model with the interval prediction model, sampling through spatial correlation and multi-confidence power interval restrictions can obtain reasonable and guiding output scenarios;

[0139] (3) Compared with the existing data-driven distributed robust optimization method, the optimization process of the embodiment of the present invention comprehensively considers the relevance and uncertainty of renewable energy, realizes the coordination between multiple devices in the system, and solves the power balance obstacle caused by the uncertainty of renewable energy;

[0140] (4) Compared with the traditional environmental economic dispatch model, the present invention adopts different control strategies to control carbon emissions and atmospheric pollutant emissions respectively, and more specifically reduces the environmental impact of the power system;

[0141] (5) It can make full use of historical experience data for modeling and prediction, and ensure the economy of the dispatch plan under the premise of stable operation of the power system, providing strong technical support for power grid regulation.

[0142] Next, the environmental economic distribution robust scheduling device considering the relevance of renewable energy proposed in accordance with an embodiment of the present invention will be described with reference to the accompanying drawings.

[0143] Figure 6 It is a block diagram of an environmental economic distribution robust scheduling device considering the relevance of renewable energy according to an embodiment of the present invention.

[0144] like Figure 6 As shown, the environmental economic distributed robust scheduling device 60 considering the relevance of renewable energy includes: a construction module 601, a prediction module 602, a clustering module 603, a distributed robust optimization module 604 and a formulation module 605.

[0145] Among them, the construction module 601 is used to establish the R-copula joint distribution function of multiple renewable energy power stations. The prediction module 602 is used to perform interval prediction on the power of each renewable energy power station to obtain a power interval with multiple confidence levels. The clustering module 603 is used to extract a preset number of renewable energy power data according to the power interval with multiple confidence levels and the R-copula joint distribution function, cluster the preset number of renewable energy power data, and generate multiple typical renewable energy output scenarios. The distributed robust optimization module 604 is used to input multiple typical renewable energy output scenarios into the target dispatching center for environmental and economic distributed robust optimization to obtain an environmental and economic distributed robust optimization model; the formulation module 605 is used to use the environmental and economic distributed robust optimization model to control carbon emissions and atmospheric pollutant emissions, so as to obtain the start-stop plan and optimal resource allocation value of the thermal power unit, and formulate a dispatching plan that meets the power balance of the power grid and is environmentally friendly according to the start-stop plan and the optimal resource allocation value.

[0146] Furthermore, in one embodiment of the present invention, the construction module 601 includes:

[0147] An acquisition unit is used to acquire historical power data of multiple renewable energy power stations, and to count historical power data of multiple wind and solar power stations to obtain edge condition distribution;

[0148] The construction unit is used to construct the R-copula joint distribution function according to the edge condition distribution to describe the spatial correlation of the outputs of multiple renewable energy power stations in the same region.

[0149] Furthermore, in one embodiment of the present invention, the prediction module 602 includes:

[0150] The prediction unit is used to predict the historical prediction results based on the historical power data of multiple wind and solar power stations based on the LSTM algorithm;

[0151] A calculation unit, used to calculate the prediction error based on the historical prediction results and the historical power data of multiple wind and solar power stations;

[0152] The generating unit is used to statistically fit the prediction error using the ABKDE method to generate interval prediction results with multiple confidence levels.

[0153] Furthermore, in one embodiment of the present invention, the generating unit includes:

[0154] The ABKDE method is used to fit the prediction errors in different power intervals, and the prediction error probability density curves in different power intervals are obtained respectively;

[0155] The prediction error probability density curves in different power intervals determine the interval prediction results with multiple confidence levels.

[0156] Furthermore, in one embodiment of the present invention, the clustering module 603 includes:

[0157] Generate random power samples uniformly in the Rvine copula joint distribution function;

[0158] Power samples of power intervals satisfying multiple confidence levels are extracted from random number power samples by an inverse transformation sampling method, and power samples of power intervals not satisfying multiple confidence levels are resampled until a preset number of samples is reached, thereby obtaining a preset number of renewable energy power data;

[0159] A fuzzy clustering algorithm is used to cluster a preset number of renewable energy power data to generate typical renewable energy output scenarios.

[0160] Further, in one embodiment of the present invention, the formulation module 605 includes:

[0161] The environmental-economic distributed robust optimization model is used to control carbon emissions and atmospheric pollutant emissions, and the column and constraint generation algorithm is used to perform distributed robust optimization on multiple typical renewable energy output scenarios until the column and constraint generation algorithm converges to obtain the start and stop plans and optimal resource allocation of thermal power units.

[0162] It should be noted that the aforementioned explanation of the embodiment of the environmental-economically robust scheduling method considering the relevance of renewable energy is also applicable to the environmental-economically robust scheduling device considering the relevance of renewable energy in this embodiment, and will not be repeated here.

[0163] The environmental economic distribution robust scheduling device considering the relevance of renewable energy proposed in the embodiment of the present invention has the following beneficial effects:

[0164] (1) The LSTM-ABKDE algorithm can be used to characterize the fluctuation range corresponding to the point prediction result under a predetermined confidence level. Compared with the traditional KDE method, the adaptive kernel density estimation (ABKDE) can use different bandwidths to complete the kernel density estimation for the power data in the window, thereby improving the accuracy and reliability of the estimation.

[0165] (2) Combining the R-vine copula model with the interval prediction model, sampling through spatial correlation and multi-confidence power interval restrictions can obtain reasonable and guiding output scenarios;

[0166] (3) Compared with the existing data-driven distributed robust optimization method, the optimization process of the embodiment of the present invention comprehensively considers the relevance and uncertainty of renewable energy, realizes the coordination between multiple devices in the system, and solves the power balance obstacle caused by the uncertainty of renewable energy;

[0167] (4) Compared with the traditional environmental economic dispatch model, the present invention adopts different control strategies to control carbon emissions and atmospheric pollutant emissions respectively, and more specifically reduces the environmental impact of the power system;

[0168] (5) It can make full use of historical experience data for modeling and prediction, and ensure the economy of the dispatch plan under the premise of stable operation of the power system, providing strong technical support for power grid regulation.

[0169] Figure 7 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. The electronic device may include:

[0170] A memory 701 , a processor 702 , and a computer program stored in the memory 701 and executable on the processor 702 .

[0171] When the processor 702 executes the program, the environmental economic distribution robust scheduling method considering the relevance of renewable energy provided in the above embodiments is implemented.

[0172] Furthermore, the electronic device further comprises:

[0173] The communication interface 703 is used for communication between the memory 701 and the processor 702 .

[0174] The memory 701 is used to store computer programs that can be executed on the processor 702 .

[0175] The memory 701 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0176] If the memory 701, the processor 702 and the communication interface 703 are implemented independently, the communication interface 703, the memory 701 and the processor 702 can be connected to each other through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 7 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0177] Optionally, in a specific implementation, if the memory 701, the processor 702 and the communication interface 703 are integrated on a chip, the memory 701, the processor 702 and the communication interface 703 can communicate with each other through an internal interface.

[0178] The processor 702 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention.

[0179] An embodiment of the present invention further provides a computer program product, which, when executed by a processor, implements the above-mentioned environmentally economically robust scheduling method considering the relevance of renewable energy.

[0180] An embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned environmentally economically distributed robust scheduling method considering the relevance of renewable energy.

[0181] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0182] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "N" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0183] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present invention belong.

[0184] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purpose of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or N wirings (electronic devices), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways as necessary and then storing it in a computer memory.

[0185] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above embodiment, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. If implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0186] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0187] In addition, each functional unit in each embodiment of the present invention may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0188] The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present invention. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present invention.

Claims

1. An environmentally economically robust scheduling method considering the relevance of renewable energy, characterized in that: The following steps are involved: Establish the R-vine copula joint distribution function of multiple renewable energy power plants; Performing interval prediction on the power of each renewable energy power station to obtain power intervals with multiple confidence levels; Extracting a preset number of renewable energy power data according to the multi-confidence power interval and the R-copula joint distribution function, clustering the preset number of renewable energy power data, and generating a plurality of typical renewable energy output scenarios; The multiple typical renewable energy output scenarios are input into the target dispatching center for environmental economic distribution robust optimization to obtain an environmental economic distribution robust optimization model, wherein the environmental economic distribution robust optimization model is modeled in two stages. In the first stage, the multiple typical renewable energy output scenarios are used as input to construct an objective function to minimize the start-up and shutdown costs, energy storage costs, carbon emission costs and system operation costs of the thermal power units, while considering the start-up and shutdown constraints of the thermal power units, the climbing constraints, the power balance constraints of the system, the line safety constraints and the atmospheric pollution environment constraints, and solving the start-up and shutdown states of the thermal power units; in the second stage, the start-up and shutdown states of the thermal power units are used as input quantities, and the 1-norm and ∞-norm are constructed based on the historical output data of renewable energy to constrain the probability of renewable energy output scenarios, and the optimal values ​​of the pre-output and resource allocation of the thermal power units are solved under the worst probability distribution conditions; The environmental economic distribution robust optimization model is used to control carbon emissions and atmospheric pollutant emissions to obtain the start-stop plan and optimal resource allocation value of the thermal power unit, and a dispatching plan that satisfies the power balance of the power grid and is environmentally friendly is formulated according to the start-stop plan and the optimal resource allocation value, wherein: The method of controlling carbon emissions and emissions of atmospheric pollutants by using the environmental economic distribution robust optimization model to obtain the start-stop plan and resource allocation optimal value of the thermal power unit includes: The environmental-economic distributed robust optimization model is used to control carbon emissions and emissions of atmospheric pollutants, and the column and constraint generation algorithm is used to perform distributed robust optimization on the multiple typical renewable energy output scenarios until the column and constraint generation algorithm converges to obtain the start and stop plan of the thermal power unit and the optimal value of resource allocation, wherein the total amount of carbon emissions is controlled, and the emission of atmospheric pollutants is controlled by diffusion concentration.

2. The environmental economic distribution robust scheduling method considering the relevance of renewable energy according to claim 1 is characterized in that: The R copula joint distribution function of establishing multiple renewable energy power stations includes: Acquire historical power data of multiple renewable energy power stations, and collect historical power data of the multiple wind and solar power stations to obtain edge condition distribution; According to the marginal condition distribution, an R-copula joint distribution function is constructed to describe the spatial correlation of outputs among multiple renewable energy power stations in the same region.

3. The environmental economic distribution robust scheduling method considering the relevance of renewable energy according to claim 2 is characterized in that: The performing interval prediction on the power of each renewable energy power station to obtain interval prediction results with multiple confidence levels includes: Based on the LSTM algorithm, historical prediction results are predicted according to the historical power data of the multiple wind and solar power stations; Calculating a prediction error based on the historical prediction results and the historical power data of the plurality of wind and solar power stations; The ABKDE method is used to statistically fit the prediction errors to generate the interval prediction results with multiple confidence levels.

4. The environmental economic distribution robust scheduling method considering the relevance of renewable energy according to claim 3 is characterized in that: The ABKDE method is used to statistically fit the prediction error to generate the interval prediction results with multiple confidence levels, including: The ABKDE method is used to fit the prediction errors in different power intervals, and the prediction error probability density curves in different power intervals are obtained respectively; The prediction error probability density curves in the different power intervals determine the multiple confidence interval prediction results.

5. The environmental economic distribution robust scheduling method considering the relevance of renewable energy according to claim 1 is characterized in that: The step of extracting a preset number of renewable energy power data according to the multi-confidence power interval and the R-copula joint distribution function, clustering the preset number of renewable energy power data, and generating a typical output scenario includes: Generating random power samples uniformly in the R-copula joint distribution function; Extracting power samples of the power interval that meets the multiple confidence levels from the random power samples by an inverse transformation sampling method, and resampling power samples of the power interval that do not meet the multiple confidence levels until a preset number of samples is reached, thereby obtaining the preset number of renewable energy power data; The preset number of renewable energy power data are clustered using a fuzzy clustering algorithm to generate the typical renewable energy output scenario.

6. An environmental economic distribution robust scheduling device considering the relevance of renewable energy, characterized in that: include: Building module for building R-vine copula joint distribution function of multiple renewable energy power plants; A prediction module, used for performing interval prediction on the power of each renewable energy power station to obtain power intervals with multiple confidence levels; A clustering module, configured to extract a preset number of renewable energy power data according to the multi-confidence power intervals and the R-copula joint distribution function, cluster the preset number of renewable energy power data, and generate a plurality of typical renewable energy output scenarios; A distributed robust optimization module is used to input the multiple typical renewable energy output scenarios into the target dispatching center for environmental and economic distributed robust optimization to obtain an environmental and economic distributed robust optimization model, wherein the environmental and economic distributed robust optimization model is modeled in two stages. In the first stage, the multiple typical renewable energy output scenarios are used as input to construct an objective function to minimize the start-up and shutdown costs, energy storage costs, carbon emission costs and system operation costs of the thermal power units, while considering the start-up and shutdown constraints of the thermal power units, the climbing constraints, the power balance constraints of the system, the line safety constraints and the atmospheric pollution environment constraints, and solve the start-up and shutdown states of the thermal power units; in the second stage, the start-up and shutdown states of the thermal power units are used as input quantities, and the 1-norm and ∞-norm are constructed based on the historical output data of renewable energy to constrain the probability of renewable energy output scenarios, and the optimal values ​​of the pre-output and resource allocation of the thermal power units are solved under the worst probability distribution conditions; A formulation module is used to control carbon emissions and atmospheric pollutant emissions by using the environmental economic analysis robust optimization model to obtain the start-stop plan and optimal resource allocation value of the thermal power unit, and formulate a dispatching plan that satisfies the power balance of the power grid and is environmentally friendly according to the start-stop plan and the optimal resource allocation value, wherein: The method of controlling carbon emissions and emissions of atmospheric pollutants by using the environmental economic distribution robust optimization model to obtain the start-stop plan and resource allocation optimal value of the thermal power unit includes: The environmental-economic distributed robust optimization model is used to control carbon emissions and emissions of atmospheric pollutants, and the column and constraint generation algorithm is used to perform distributed robust optimization on the multiple typical renewable energy output scenarios until the column and constraint generation algorithm converges to obtain the start and stop plan of the thermal power unit and the optimal value of resource allocation, wherein the total amount of carbon emissions is controlled, and the emission of atmospheric pollutants is controlled by diffusion concentration.

7. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the environmentally economically distributed robust scheduling method considering the relevance of renewable energy as described in any one of claims 1 to 5.

8. A computer program product, characterized in that When the computer program / instructions are executed by a processor, the environmentally economically distributed robust scheduling method considering the relevance of renewable energy as described in any one of claims 1 to 5 is implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement the environmentally economically distributed robust scheduling method considering the relevance of renewable energy as described in any one of claims 1 to 5.

Images