A method, device, equipment and medium for evaluating the probability density of wind power

Abstract

The present invention discloses a probability density evaluation method, device, equipment and medium for wind power. By acquiring the output sample data and upper and lower limit data of the wind power of the wind farm, a probability density function estimation model based on the logarithmic kernel differential homeomorphism estimation principle is established; according to the difference test, chi-square test and KS test between the probability model and the data observation distribution, a multi-objective optimization calculation model of the window width of the probability density function estimation model is established; the multi-objective optimization calculation model is solved to obtain the optimal solution set of the window width; the final probability density function of the wind power is solved, and the wind power of the wind farm is evaluated according to the final probability density function. The provided method can take into account that wind power is a bounded random variable, can coordinate multiple evaluation indicators at the same time, and use the wind power probability density function model obtained by the provided method to evaluate the wind power output of the wind farm, providing technical support for the operation and planning of large-scale wind power systems.

Description

Technical Field

[0001] The present invention relates to the technical field of power system operation and planning, and in particular to a method, device, equipment and medium for evaluating the probability density of wind power. Background Art

[0002] As the scale of wind farms connected to the power system becomes larger and larger, the random characteristics of wind farm output can no longer be ignored in system analysis. The probability density function of wind farm output can accurately estimate the wind farm output. Therefore, studying the calculation method of the probability density function of wind farm output has important practical significance.

[0003] There are many methods for estimating the probability density function of wind farm output, such as parameter estimation and non-parameter estimation. Currently, the kernel density estimation method among non-parameter estimation methods is the mainstream method. However, the kernel density estimation methods in the prior art do not take into account the fact that wind power is a bounded random variable. When the existing kernel density estimation methods are used for probability density estimation of bounded interval variables, the boundary of the interval often has boundary deviation problems due to the lack of good convergence, thereby affecting the accuracy of probability density estimation. Summary of the invention

[0004] The embodiments of the present invention provide a method, device, equipment and medium for evaluating the probability density of wind power, which take into account the boundedness of the random variable of wind power, reduce the boundary deviation of the probability density function, and improve the accuracy of probability density estimation.

[0005] An embodiment of the present invention provides a method for evaluating the probability density of wind power, the method comprising:

[0006] Obtaining output sample data of a wind farm and upper and lower limit data of wind power of the wind farm, and establishing a probability density function estimation model of wind power of the wind farm based on the logarithmic kernel differential homeomorphism estimation principle;

[0007] According to the difference test, chi-square test and KS test between the probability model and the data observation distribution, a multi-objective optimization calculation model for the window width of the probability density function estimation model is established;

[0008] Using a multi-objective particle swarm algorithm to solve the multi-objective optimization calculation model, and obtain the window width optimal solution set of the probability density function estimation model;

[0009] A final probability density function is solved according to the window width optimal solution set and the probability density function estimation model, and the wind power of the wind farm is evaluated according to the final probability density function.

[0010] Preferably, the acquiring of sample data of wind farm output and upper and lower limit data of wind power of the wind farm, and establishing a probability density function estimation model of the wind farm specifically includes:

[0011] Obtain the sample data P of the wind farm output w1 , P w2 ,…,P wn And the wind power upper and lower limit data [P min ,P max ];

[0012] According to the logarithmic kernel diffeomorphism estimation principle, the probability density estimation model of the wind power of the wind farm is calculated:

[0013]

[0014]

[0015]

[0016] Where P is the actual wind power output of the wind farm, P∈[P min ,P max ], f d,n (P) is the probability density function of the wind power of the wind farm, F(P) is the probability distribution function of the wind power of the wind farm, h d is the window width obtained after logarithmic diffeomorphic transformation of the output sample data of the wind farm, K() is the kernel function, P wi ∈[P min ,P max ], i=1,2,…,n.

[0017] Preferably, the multi-objective optimization solution model is established by using the difference test between the probability model and the data observation distribution, the chi-square test and the KS test, specifically including:

[0018] The output sample data P w1 , P w2 ,…,P wn Divide into m groups of non-intersecting data sets, and calculate the chi-square test statistic χ of the chi-square test 2 ;

[0019] The sample data sample P w1 , P w2 ,…,P wn Sort from small to large and calculate the KS test statistic D of the KS test n ;

[0020] The wind power upper and lower limit data [P min ,P max] is divided into M intervals, and the output sample data P appearing in each interval is counted w1 , P w2 ,…,P wn The power histogram is used to calculate the difference test statistic MSR for the difference test between the probability model and the data observation distribution;

[0021] Establish the multi-objective optimization solution model of the window width: objective function Objective function f2 = minD n , objective function f3 = minMSR and optimization variable h dx ;

[0022] in, k i is the observed frequency of the ith set in m sets of non-intersecting data sets, p i is the theoretical probability value of the i-th set calculated by the probability density function, is a chi-square distribution with m-1 degrees of freedom. When the confidence level is given as α, for The 1-α quantile of

[0023] P w ' i The sample data P w1 , P w2 ,…,P wn The i-th sample after sorting from small to large;

[0024] y i is the probability of the wind power of the wind farm in the ith interval calculated by the probability density function, x i is the probability of the wind power of the wind farm being in the i-th interval obtained by counting the power histogram;

[0025] h dx ∈[h min ,h max ],h min 、h max are the upper and lower limits of the preset optimization variables respectively.

[0026] Preferably, the multi-objective particle swarm algorithm is used to solve the multi-objective optimization calculation model to obtain the window width optimal solution set of the probability density function estimation model, specifically including:

[0027] Establish a particle group, the particle group includes: a main particle group and three from the particle swarm

[0028] Establish an external reserve set, initialize the particle group and the speed of each particle in the particle group, and solve the optimization variable h of the multi-objective optimization model according to the optimization variable h of the multi-objective optimization model. dx Each particle is randomly generated according to the constraints;

[0029] According to the particle optimization update formula, the particles of the particle group are updated once, and the non-inferior extreme value particles obtained from the three slave particle groups are added to the main particle group, the three non-inferior particles from the slave particle group are added to the external reserve group, the non-inferior particles obtained from the main particle group are added to the external reserve set, the particle with the smallest value of the corresponding objective function f1 in the external reserve group is used as the global guiding particle of the slave particle group S1, the particle with the smallest value of the corresponding objective function f2 in the external reserve group is used as the global guiding particle of the slave particle group S2, the particle with the smallest value of the corresponding objective function f3 in the external reserve group is used as the global guiding particle of the slave particle group S3, and the particle with the highest sparsity in the external reserve group is used as the global guiding particle of the main particle group S0, completing an update optimization process;

[0030] Repeat the particle group update optimization process until the number of iterations reaches a preset maximum value, and use the latest external reserve set as the window width optimal solution set of the probability density function estimation model;

[0031] The definition of sparsity is as follows: for particle i in the external reserve group, the sparsity DP of particle i is measured as i :

[0032]

[0033] N k is the number of particles in the external reserve set after the kth update optimization, f1 j , The corresponding objective function f1, f2, and f3 values ​​of particle j in the external reserve group. i The larger the value, the sparser the distribution of particles i is. i The largest one serves as the global guiding particle of the main particle group S0.

[0034] Where N' is the number of particles in the master particle group, N is the number of particles in each slave particle group, and the optimization update formula is: and and The particle swarm S at the k+1th and kth iterations are j The search speed of particle i, w j is the particle swarm S j The inertia weight, and The particle group S j The learning factor, r1j and The particle group S j A random number that follows a uniform distribution on [0,1], and The particle group S j The current value of particle i at the kth iteration, the value of the individual local guiding particle, and the value of the global guiding particle.

[0035] Preferably, solving a final probability density function according to the window width optimal solution set and the probability density function estimation model, and evaluating the wind power of the wind farm according to the final probability density function specifically includes:

[0036] Selecting any solution in the window width optimal solution set as the final solution of the window width, substituting the optimal solution of the window width into the probability density function estimation model to obtain the final probability density function of the wind farm output;

[0037] And the wind power of the wind farm is evaluated according to the final probability density function.

[0038] The embodiment of the present invention also provides a probability density evaluation device for wind power, the device comprising: a density model module, an optimization model module, a solution module and an evaluation module;

[0039] The density model module is used to obtain the output sample data of the wind farm and the upper and lower limit data of the wind power of the wind farm, and establish a probability density function estimation model of the wind power of the wind farm based on the logarithmic kernel differential homeomorphism estimation principle;

[0040] The optimization model module is used to establish a multi-objective optimization calculation model for the window width of the probability density function estimation model according to the difference test, chi-square test and KS test between the probability model and the data observation distribution;

[0041] The solution module is used to solve the multi-objective optimization calculation model using a multi-objective particle swarm algorithm to obtain a window width optimal solution set of the probability density function estimation model;

[0042] The evaluation module is used to solve a final probability density function according to the window width optimal solution set and the probability density function estimation model, and evaluate the wind power of the wind farm according to the final probability density function.

[0043] As a preferred embodiment, the density model module is specifically used for:

[0044] Obtain the sample data P of the wind farm output w1 , P w2 ,…,P wn And the wind power upper and lower limit data [P min ,Pmax ];

[0045] According to the logarithmic kernel diffeomorphism estimation principle, the probability density estimation model of the wind power of the wind farm is calculated:

[0046]

[0047]

[0048]

[0049] Where P is the actual wind power output of the wind farm, P∈[P min ,P max ], f d,n (P) is the probability density function of the wind power of the wind farm, F(P) is the probability distribution function of the wind power of the wind farm, h d is the window width obtained after logarithmic diffeomorphic transformation of the output sample data of the wind farm, K() is the kernel function, P wi ∈[P min ,P max ], i=1,2,…,n.

[0050] As a preferred embodiment, the optimization model module is specifically used for:

[0051] The output sample data P w1 , P w2 ,…,P wn Divide into m groups of non-intersecting data sets, and calculate the chi-square test statistic χ of the chi-square test 2 ;

[0052] The sample data sample P w1 , P w2 ,…,P wn Sort from small to large and calculate the KS test statistic D of the KS test n ;

[0053] The wind power upper and lower limit data [P min ,P max ] is divided into M intervals, and the output sample data P appearing in each interval is counted w1 , P w2 ,…,P wn The power histogram is used to calculate the difference test statistic MSR for the difference test between the probability model and the data observation distribution;

[0054] Establish the multi-objective optimization solution model of the window width: objective function Objective function f2 = minD n, objective function f3 = minMSR and optimization variable h dx ;

[0055] in, k i is the observed frequency of the ith set in m sets of non-intersecting data sets, p i is the theoretical probability value of the i-th set calculated by the probability density function, is a chi-square distribution with m-1 degrees of freedom. When the confidence level is given as α, for The 1-α quantile of

[0056] P′ wi The sample data P w1 , P w2 ,…,P wn The i-th sample after sorting from small to large;

[0057] y i is the probability of the wind power of the wind farm in the ith interval calculated by the probability density function, x i is the probability of the wind power of the wind farm being in the i-th interval obtained by counting the power histogram;

[0058] h dx ∈[h min ,h max ],h min 、h max are the upper and lower limits of the preset optimization variables respectively.

[0059] As a preferred embodiment, the solution module is specifically used for:

[0060] Establish a particle group, the particle group includes: a main particle group and three from the particle swarm

[0061] Establish an external reserve set, initialize the particle group and the speed of each particle in the particle group, and solve the optimization variable h of the multi-objective optimization model according to the optimization variable h of the multi-objective optimization model. dx Each particle is randomly generated according to the constraints;

[0062] According to the particle optimization update formula, the particles of the particle group are updated once, and the non-inferior extreme value particles obtained from the three slave particle groups are added to the main particle group, the three non-inferior particles from the slave particle group are added to the external reserve group, the non-inferior particles obtained from the main particle group are added to the external reserve set, the particle with the smallest value of the corresponding objective function f1 in the external reserve group is used as the global guiding particle of the slave particle group S1, the particle with the smallest value of the corresponding objective function f2 in the external reserve group is used as the global guiding particle of the slave particle group S2, the particle with the smallest value of the corresponding objective function f3 in the external reserve group is used as the global guiding particle of the slave particle group S3, and the particle with the highest sparsity in the external reserve group is used as the global guiding particle of the main particle group S0, completing an update optimization process;

[0063] Repeat the particle group update optimization process until the number of iterations reaches a preset maximum value, and use the latest external reserve set as the window width optimal solution set of the probability density function estimation model;

[0064] The definition of sparsity is as follows: For particle i in the external reserve group, the sparsity DP of particle i is measured as i :

[0065]

[0066] N k is the number of particles in the external reserve set after the kth update optimization, f1 j , The corresponding objective function f1, f2, and f3 values ​​of particle j in the external reserve group. i The larger the value, the sparser the distribution of particles i is. i The largest one serves as the global guiding particle of the main particle group S0;

[0067] N' is the number of particles in the master particle swarm, N is the number of particles in each slave particle swarm, and the optimization update formula is: and and The particle swarm S at the k+1th and kth iterations are j The search speed of particle i, w j is the particle swarm S j The inertia weight, and The particle group S j The learning factor, r1 j and The particle group S j A random number that follows a uniform distribution on [0,1], and The particle group Sj The current value of particle i at the kth iteration, the value of the individual local guiding particle, and the value of the global guiding particle.

[0068] As a preferred embodiment, the evaluation module is specifically used for:

[0069] Selecting any solution in the window width optimal solution set as the final solution of the window width, substituting the optimal solution of the window width into the probability density function estimation model to obtain the final probability density function of the wind farm output;

[0070] And the wind power of the wind farm is evaluated according to the final probability density function.

[0071] Another embodiment of the present invention provides a storage medium, wherein the computer-readable storage medium includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the probability density assessment method of wind power described in the above-mentioned embodiment of the invention.

[0072] Another embodiment of the present invention provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and when the processor executes the computer program, the probability density assessment method of wind power described in the above-mentioned embodiment of the invention is implemented.

[0073] The present invention provides a method, device, equipment and medium for evaluating the probability density of wind power. By acquiring the output sample data of the wind farm and the upper and lower limit data of the wind power of the wind farm, a probability density function estimation model of the wind power of the wind farm based on the logarithmic kernel differential homeomorphism estimation principle is established; according to the difference test, chi-square test and KS test between the probability model and the data observation distribution, a multi-objective optimization calculation model of the window width of the probability density function estimation model is established; the multi-objective optimization calculation model is solved to obtain the optimal solution set of the window width of the probability density function estimation model; the final probability density function is solved, and the wind power of the wind farm is evaluated according to the final probability density function. It can take into account that the suitable wind power is a bounded random variable, and evaluate the wind power of the wind farm output according to the wind power probability density function model, providing technical support for the operation and planning of large-scale wind power systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Figure 1 It is a flow chart of a method for evaluating the probability density of wind power provided by an embodiment of the present invention;

[0075] Figure 2 is a schematic diagram of a particle swarm optimization process provided by an embodiment of the present invention;

[0076] Figure 3It is a structural schematic diagram of a wind power probability density assessment device provided by an embodiment of the present invention;

[0077] Figure 4 It is a structural diagram of a terminal device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0078] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0079] The embodiment of the present invention provides a probability density evaluation method for wind power, see Figure 1 , is a flow chart of a probability density evaluation method for wind power provided by an embodiment of the present invention, wherein steps S101 to S104 of the method are as follows:

[0080] S101, obtaining output sample data of a wind farm and upper and lower limit data of wind power of the wind farm, and establishing a probability density function estimation model of wind power of the wind farm based on the logarithmic kernel diffeomorphism estimation principle;

[0081] S102, establishing a multi-objective optimization calculation model for the window width of the probability density function estimation model according to a difference test, a chi-square test, and a KS test between the probability model and the data observation distribution;

[0082] S103, solving the multi-objective optimization calculation model using a multi-objective particle swarm algorithm to obtain a window width optimal solution set of the probability density function estimation model;

[0083] S104, solving a final probability density function according to the window width optimal solution set and the probability density function estimation model, and evaluating the wind power of the wind farm according to the final probability density function.

[0084] In the specific implementation of this embodiment, the output sample data of the wind farm and the upper and lower limit data of the wind power of the wind farm are obtained, and a probability density function estimation model of the wind power of the wind farm based on the logarithmic kernel differential homeomorphism estimation principle is established;

[0085] A multi-objective optimization calculation model with three objectives is established by using chi-square test, KS test and difference test between probability model and data observation distribution (statistical histogram) to solve the optimal window width.

[0086] Using a multi-objective particle swarm algorithm to solve the multi-objective optimization calculation model, and obtain the optimal solution set of the window width of the Pareto of the window width;

[0087] A final probability density function is solved according to the window width optimal solution set and the probability density function estimation model, and the wind power of the wind farm is evaluated according to the final probability density function.

[0088] By considering the boundedness of the random variables of the wind farm output and calculating the probability density function of the wind power of the wind farm, the output wind power of the wind farm can be accurately evaluated, providing technical support for the operation and planning of large-scale wind power systems.

[0089] In another embodiment provided by the present invention, the step S101 specifically includes:

[0090] Obtain the sample data P of the wind farm output w1 , P w2 ,…,P wn And the wind power upper and lower limit data [P min ,P max ];

[0091] According to the logarithmic kernel diffeomorphism estimation principle, the probability density estimation model of the wind power of the wind farm is calculated:

[0092]

[0093]

[0094]

[0095] Where P is the actual wind power output of the wind farm, P∈[P min ,P max ], f d,n (P) is the probability density function of the wind power of the wind farm, F(P) is the probability distribution function of the wind power of the wind farm, h d is the window width obtained after logarithmic diffeomorphic transformation of the output sample data of the wind farm, K() is the kernel function, P wi ∈[P min ,P max ], i=1,2,…,n.

[0096] In the specific implementation of this embodiment, the wind power upper and lower limit data of the wind farm [P min ,P max ], the output sample data of the wind farm is P w1 , P w2 ,…,P wn , the output sample data are all within the upper and lower limit data range of wind power, that is, P wi ∈[P min ,P max], i=1,2,…,n;

[0097] P is the actual output of the wind farm, P∈[P min ,P max ], according to the logarithmic kernel diffeomorphism estimation principle, the probability density function of wind power in the wind farm is f d,n (P) and probability distribution function F d (P) is:

[0098]

[0099]

[0100]

[0101] in, is a function The first derivative of d is the initial known sample P wi (i=1,2,…,n) is the window width obtained after logarithmic diffeomorphism transformation, and K() is the kernel function.

[0102] It should be noted that the kernel function K() includes many types, such as Gaussian kernel function, Epanechnikov kernel function, triangular kernel function, etc.

[0103] Take the Gaussian kernel function as an example to illustrate the specific application of the kernel function.

[0104] The wind power is converted into a bounded random variable using logarithmic diffeomorphism transformation, and a probability density function model of wind power after diffeomorphism transformation is established. Considering the boundedness of wind power conversion, the wind power estimation model is more accurate.

[0105] In another embodiment provided by the present invention, the step S102 specifically includes:

[0106] The output sample data P w1 , P w2 ,…,P wn Divide into m groups of non-intersecting data sets, and calculate the chi-square test statistic χ of the chi-square test 2 ;

[0107] The sample data sample P w1 , P w2 ,…,P wn Sort from small to large and calculate the KS test statistic D of the KS test n ;

[0108] The wind power upper and lower limit data [P min ,P max] is divided into M intervals, and the output sample data P appearing in each interval is counted w1 , P w2 ,…,P wn The power histogram is used to calculate the difference test statistic MSR for the difference test between the probability model and the data observation distribution;

[0109] Establish the multi-objective optimization solution model of the window width: objective function Objective function f2 = minD n , objective function f3 = minMSR and optimization variable h dx ;

[0110] in, k i is the observed frequency of the ith set in m sets of non-intersecting data sets, p i is the theoretical probability value of the i-th set calculated by the probability density function, is a chi-square distribution with m-1 degrees of freedom. When the confidence level is given as α, for The 1-α quantile of

[0111] P′ wi The sample data P w1 , P w2 ,…,P wn The i-th sample after sorting from small to large;

[0112] y i is the probability of the wind power of the wind farm in the ith interval calculated by the probability density function, x i is the probability of the wind power of the wind farm being in the i-th interval obtained by counting the power histogram;

[0113] h dx ∈[h min ,h max ],h min 、h max are the upper and lower limits of the preset optimization variables respectively.

[0114] In the specific implementation of this embodiment, the chi-square test, KS test and the difference test between the probability model and the data observation distribution are selected to establish a multi-objective optimization solution model. The specific process is:

[0115] Suppose the number of wind farm output samples is n and it is P w1 , P w2 ,…,P wn, divide the sample data into m groups of non-intersecting data, and calculate the chi-square test statistic of the chi-square test Among them, k i is the observed frequency of the ith set in m sets of non-intersecting data sets, p i is the theoretical probability value of the i-th set calculated by the probability density function, which can be calculated by the probability distribution function and the probability density function of the probability density estimation model in the previous embodiment; is a chi-square distribution with m-1 degrees of freedom. When the confidence level is given as α, for The 1-α quantile of , the value can be obtained by looking up the table, and

[0116] Calculate the KS test statistic for the KS test Where P w ' i For sample P w1 , P w2 ,…,P wn The i-th sample after sorting from small to large, D n The smaller the better.

[0117] The wind power upper and lower limit data [P min ,P max ] is divided into M intervals, and the output sample data P appearing in each interval is counted w1 , P w2 ,…,P wn The power histogram of the probability model is used to calculate the difference test statistic between the probability model and the data observation distribution (statistical histogram) Among them, y i is the probability of the wind power of the wind farm in the ith interval calculated by the probability density function, x i is the probability of the wind power of the wind farm being in the i-th interval calculated by using the power histogram; the smaller the MSR, the better.

[0118] A multi-objective optimization solution model for the window width is established:

[0119] Objective Function

[0120] Objective function f2 = minD n ;

[0121] Objective function f3 = minMSR;

[0122] Optimize variable h dx ∈[h min ,h max ];

[0123] h min 、h max are the upper and lower limits of the preset optimization variables respectively.

[0124] The kernel estimation model of the wind power probability density function after diffeomorphism transformation is established, and then the chi-square test (χ 2 A multi-objective optimization solution model is established by performing a test on the difference between the probability model and the data observation distribution (statistical histogram), a KS test, and a test on the difference between the probability model and the data observation distribution (statistical histogram) to solve the window width of the probability density function estimation model of the wind power of the wind farm based on the logarithmic kernel differential homeomorphism estimation principle.

[0125] In another embodiment provided by the present invention, the step S103 specifically includes:

[0126] Establish a particle group, the particle group includes: a main particle group and three from the particle swarm

[0127] Establish an external reserve set, initialize the particle group and the speed of each particle in the particle group, and solve the optimization variable h of the multi-objective optimization model according to the optimization variable h of the multi-objective optimization model. dx Each particle is randomly generated according to the constraints;

[0128] According to the particle optimization update formula, the particles of the particle group are updated once, and the non-inferior extreme value particles obtained from the three slave particle groups are added to the main particle group, the three non-inferior particles from the slave particle group are added to the external reserve group, the non-inferior particles obtained from the main particle group are added to the external reserve set, the particle with the smallest value of the corresponding objective function f1 in the external reserve group is used as the global guiding particle of the slave particle group S1, the particle with the smallest value of the corresponding objective function f2 in the external reserve group is used as the global guiding particle of the slave particle group S2, the particle with the smallest value of the corresponding objective function f3 in the external reserve group is used as the global guiding particle of the slave particle group S3, and the particle with the highest sparsity in the external reserve group is used as the global guiding particle of the main particle group S0, completing an update optimization process;

[0129] Repeat the particle group update optimization process until the number of iterations reaches a preset maximum value, and use the latest external reserve set as the window width optimal solution set of the probability density function estimation model;

[0130] The definition of sparsity is as follows: for particle i in the external reserve group, the sparsity DP of particle i is measured as i :

[0131]

[0132] N kis the number of particles in the external reserve set after the kth update optimization, f1 j , The corresponding objective function f1, f2, and f3 values ​​of particle j in the external reserve group. i The larger the value, the sparser the distribution of particles i is. i The largest one serves as the global guiding particle of the main particle group S0;

[0133] N' is the number of particles in the master particle swarm, N is the number of particles in each slave particle swarm, and the optimization update formula is: and and The particle swarm S at the k+1th and kth iterations are j The search speed of particle i, w j is the particle swarm S j The inertia weight, and The particle group S j The learning factor, r1 j and The particle group S j A random number that follows a uniform distribution on [0,1], and The particle group S j The current value of particle i at the kth iteration, the value of the individual local guiding particle, and the value of the global guiding particle.

[0134] When this embodiment is implemented, the specific process of solving the multi-objective optimization calculation model is as follows:

[0135] Establish a particle group, which includes 1 main particle group and 3 slave particle groups. The main particle group has N' particles, and each slave particle group has N particles, which are respectively recorded as: main particle group From particle swarm

[0136] Establish an external reserve set, initialize the particle group and the speed of each particle in the particle group, and solve the optimization variable h of the multi-objective optimization model according to the optimization variable h of the multi-objective optimization model. dx Each particle is randomly generated according to the constraints, and the external reserve set is used to record the non-inferior solutions searched;

[0137] Particle Swarm j The particles are evolved and updated according to the optimization update formula, and the optimization update formula is specifically:

[0138]

[0139]

[0140] Where j = 0, 1, 2, 3; and The particle swarm S at the k+1th and kth iterations are j The search speed of particle i, w j is the particle swarm S j The inertia weight, and The particle group S j The learning factor, r1 j and The particle group S j A random number that follows a uniform distribution on [0,1], and The particle group S j The current value of particle i at the kth iteration, the value of the individual local guiding particle, and the value of the global guiding particle.

[0141] After updating the particles in each particle group, each particle group needs to be optimized. Figure 2 , is a schematic diagram of the particle swarm optimization process provided by an embodiment of the present invention; the master particle swarm S0 is optimized by using the three objective functions of the multi-objective optimization solution model, the slave particle swarm S1 is optimized by using the objective function f1 of the multi-objective optimization solution model, the slave particle swarm S2 is optimized by using the objective function f2 of the multi-objective optimization solution model, and the slave particle swarm S3 is optimized by using the objective function f3 of the multi-objective optimization solution model;

[0142] Each time the particles of the particle group are updated according to the particle optimization update formula, the particles of each particle group are optimized, specifically:

[0143] The non-inferior extreme value particles obtained from the three slave particle groups are added to the main particle group, the three non-inferior particles from the slave particle group are added to the external reserve group, the non-inferior particles obtained from the main particle group are added to the external reserve set, the particle with the smallest value of the corresponding objective function f1 in the external reserve group is used as the global guiding particle of the slave particle group S1, the particle with the smallest value of the corresponding objective function f2 in the external reserve group is used as the global guiding particle of the slave particle group S2, the particle with the smallest value of the corresponding objective function f3 in the external reserve group is used as the global guiding particle of the slave particle group S3, and the particle with the highest sparsity in the external reserve group is used as the global guiding particle of the main particle group S0, completing an update optimization process;

[0144] The definition of sparsity is as follows:

[0145] For particle i in the external reserve group, the sparsity of particle i can be obtained by the following formula:

[0146]

[0147] N k is the number of particles in the external reserve set after the kth update optimization, f1 j , The corresponding objective function f1, f2, and f3 values ​​of particle j in the external reserve group. i The larger the value, the sparser the distribution of particles i is. i The largest one serves as the global guiding particle of the main particle group S0.

[0148] Repeat the particle group update optimization process until the number of iterations reaches a preset maximum value, and use the latest external reserve set as the window width optimal solution set of the probability density function estimation model;

[0149] In another embodiment provided by the present invention, the step S104 specifically includes:

[0150] Selecting any solution in the window width optimal solution set as the final solution of the window width, substituting the optimal solution of the window width into the probability density function estimation model to obtain the final probability density function of the wind farm output;

[0151] And the wind power of the wind farm is evaluated according to the final probability density function.

[0152] In the specific implementation of this embodiment, the solution set in the external reserve set is the optimal solution set for the window width, that is, the final candidate window width set, and one of them is selected as the final window width h according to actual needs. d The value of the window width h d Substituting the probability density function estimation model into the probability density function, a final probability density function of the wind farm output is obtained; and the wind power of the wind farm is evaluated according to the final probability density function.

[0153] The embodiment of the present invention provides a probability density evaluation method for wind power. By acquiring the output sample data of the wind farm and the upper and lower limit data of the wind power of the wind farm, a probability density function estimation model of the wind power of the wind farm based on the logarithmic kernel differential homeomorphism estimation principle is established; according to the difference test, chi-square test and KS test between the probability model and the data observation distribution, a multi-objective optimization calculation model of the window width of the probability density function estimation model is established; the multi-objective optimization calculation model is solved to obtain the window width optimal solution set of the probability density function estimation model; the final probability density function is solved, and the wind power of the wind farm is evaluated according to the final probability density function. It can take into account that the suitable wind power is a bounded random variable, and evaluate the wind power of the wind farm output according to the wind power probability density function model, providing technical support for the operation and planning of large-scale wind power systems.

[0154] The embodiment of the present invention also provides a wind power probability density evaluation device, see Figure 3 , is a schematic diagram of the structure of a probability density evaluation device for wind power provided by an embodiment of the present invention, the device comprising: a density model module, an optimization model module, a solution module and an evaluation module;

[0155] The density model module is used to obtain the output sample data of the wind farm and the upper and lower limit data of the wind power of the wind farm, and establish a probability density function estimation model of the wind power of the wind farm based on the logarithmic kernel differential homeomorphism estimation principle;

[0156] The optimization model module is used to establish a multi-objective optimization calculation model for the window width of the probability density function estimation model according to the difference test, chi-square test and KS test between the probability model and the data observation distribution;

[0157] The solution module is used to solve the multi-objective optimization calculation model using a multi-objective particle swarm algorithm to obtain a window width optimal solution set of the probability density function estimation model;

[0158] The evaluation module is used to solve a final probability density function according to the window width optimal solution set and the probability density function estimation model, and evaluate the wind power of the wind farm according to the final probability density function.

[0159] In another embodiment provided by the present invention, the density model module is specifically used for:

[0160] Obtain the sample data P of the wind farm output w1 , P w2 ,…,P wn And the wind power upper and lower limit data [P min ,P max ];

[0161] According to the logarithmic kernel diffeomorphism estimation principle, the probability density estimation model of the wind power of the wind farm is calculated:

[0162]

[0163]

[0164]

[0165] Where P is the actual wind power output of the wind farm, P∈[P min ,P max ], f d,n (P) is the probability density function of the wind power of the wind farm, F(P) is the probability distribution function of the wind power of the wind farm, hd is the window width obtained after logarithmic diffeomorphic transformation of the output sample data of the wind farm, K() is the kernel function, P wi ∈[P min ,P max ], i=1,2,…,n.

[0166] In another embodiment provided by the present invention, the optimization model module is specifically used for:

[0167] The output sample data P w1 , P w2 ,…,P wn Divide into m groups of non-intersecting data sets, and calculate the chi-square test statistic χ of the chi-square test 2 ;

[0168] The sample data sample P w1 , P w2 ,…,P wn Sort from small to large and calculate the KS test statistic D of the KS test n ;

[0169] The wind power upper and lower limit data [P min ,P max ] is divided into M intervals, and the output sample data P appearing in each interval is counted w1 , P w2 ,…,P wn The power histogram is used to calculate the difference test statistic MSR for the difference test between the probability model and the data observation distribution;

[0170] Establish the multi-objective optimization solution model of the window width: objective function Objective function f2 = minD n , objective function f3 = minMSR and optimization variable h dx ;

[0171] in, k i is the observed frequency of the ith set in m sets of non-intersecting data sets, p i is the theoretical probability value of the i-th set calculated by the probability density function, is a chi-square distribution with m-1 degrees of freedom. When the confidence level is given as α, for The 1-α quantile of

[0172] P w ' i The sample data P w1 , P w2,…,P wn The i-th sample after sorting from small to large;

[0173] y i is the probability of the wind power of the wind farm in the ith interval calculated by the probability density function, x i is the probability of the wind power of the wind farm being in the i-th interval obtained by counting the power histogram;

[0174] h dx ∈[h min ,h max ],h min 、h max are the upper and lower limits of the preset optimization variables respectively.

[0175] In another embodiment provided by the present invention, the solution module is specifically used for:

[0176] Establish a particle group, the particle group includes: a main particle group and three from the particle swarm

[0177] Establish an external reserve set, initialize the particle group and the speed of each particle in the particle group, and solve the optimization variable h of the multi-objective optimization model according to the optimization variable h of the multi-objective optimization model. dx Each particle is randomly generated according to the constraints;

[0178] According to the particle optimization update formula, the particles of the particle group are updated once, and the non-inferior extreme value particles obtained from the three slave particle groups are added to the main particle group, the three non-inferior particles from the slave particle group are added to the external reserve group, the non-inferior particles obtained from the main particle group are added to the external reserve set, the particle with the smallest value of the corresponding objective function f1 in the external reserve group is used as the global guiding particle of the slave particle group S1, the particle with the smallest value of the corresponding objective function f2 in the external reserve group is used as the global guiding particle of the slave particle group S2, the particle with the smallest value of the corresponding objective function f3 in the external reserve group is used as the global guiding particle of the slave particle group S3, and the particle with the highest sparsity in the external reserve group is used as the global guiding particle of the main particle group S0, completing an update optimization process;

[0179] Repeat the particle group update optimization process until the number of iterations reaches a preset maximum value, and use the latest external reserve set as the window width optimal solution set of the probability density function estimation model;

[0180] The definition of sparsity is as follows: for particle i in the external reserve group, the sparsity DP of particle i is measured as i :

[0181]

[0182] N k is the number of particles in the external reserve set after the kth update optimization, f1 j , The corresponding objective function f1, f2, and f3 values ​​of particle j in the external reserve group. i The larger the value, the sparser the distribution of particles i is. i The largest one serves as the global guiding particle of the main particle group S0;

[0183] Where N' is the number of particles in the master particle group, N is the number of particles in each slave particle group, and the optimization update formula is: and and The particle swarm S at the k+1th and kth iterations are j The search speed of particle i, w j is the particle swarm S j The inertia weight, and The particle group S j The learning factor, r1 j and The particle group S j A random number that follows a uniform distribution on [0,1], and The particle group S j The current value of particle i at the kth iteration, the value of the individual local guiding particle, and the value of the global guiding particle.

[0184] In another embodiment provided by the present invention, the evaluation module is specifically used for:

[0185] Selecting any solution in the window width optimal solution set as the final solution of the window width, substituting the optimal solution of the window width into the probability density function estimation model to obtain the final probability density function of the wind farm output;

[0186] And the wind power of the wind farm is evaluated according to the final probability density function.

[0187] It should be noted that, in this embodiment, the specific functions of each module have been specifically described in the above embodiments and will not be repeated here.

[0188] See also Figure 4, is a schematic diagram of the structure of a terminal device provided in an embodiment of the present invention. The terminal device of this embodiment includes: a processor, a memory, and a computer program stored in the memory and executable on the processor, such as a probability density assessment program for wind power. When the processor executes the computer program, the steps in the above-mentioned various wind power probability density assessment method embodiments are implemented, such as Figure 1 The steps shown in the method are to evaluate the probability density of wind power. Alternatively, when the processor executes the computer program, the functions of the modules in the embodiments of the above-mentioned wind power probability density evaluation device are realized.

[0189] Exemplarily, the computer program may be divided into one or more modules / units, and the one or more modules / units are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of completing specific functions, and the instruction segments are used to describe the execution process of the computer program in the terminal device. For example, the computer program may be divided into a density model module, an optimization model module, a solution module, and an evaluation module, and the specific functions of each module are not described in detail herein.

[0190] The terminal device may be a computing device such as a desktop computer, a notebook, a PDA, and a cloud server. The terminal device may include, but is not limited to, a processor and a memory. Those skilled in the art will appreciate that the schematic diagram is only an example of a terminal device and does not constitute a limitation on the terminal device. The terminal device may include more or fewer components than shown in the diagram, or may combine certain components, or different components. For example, the terminal device may also include an input / output device, a network access device, a bus, etc.

[0191] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the terminal device, and uses various interfaces and lines to connect various parts of the entire terminal device.

[0192] The memory can be used to store the computer program and / or module, and the processor realizes various functions of the terminal device by running or executing the computer program and / or module stored in the memory, and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (SmartMedia Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0193] Wherein, if the module / unit integrated in the terminal device is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Wherein, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0194] It should be noted that the device embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the accompanying drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines. A person of ordinary skill in the art may understand and implement it without paying any creative effort.

[0195] The present invention provides a method, device, equipment and medium for evaluating the probability density of wind power. By acquiring the output sample data of the wind farm and the upper and lower limit data of the wind power of the wind farm, a probability density function estimation model of the wind power of the wind farm based on the logarithmic kernel differential homeomorphism estimation principle is established; according to the difference test, chi-square test and KS test between the probability model and the data observation distribution, a multi-objective optimization calculation model of the window width of the probability density function estimation model is established; the multi-objective optimization calculation model is solved to obtain the optimal solution set of the window width of the probability density function estimation model; the final probability density function is solved, and the wind power of the wind farm is evaluated according to the final probability density function. It can take into account that the suitable wind power is a bounded random variable, and evaluate the wind power of the wind farm output according to the wind power probability density function model, providing technical support for the operation and planning of large-scale wind power systems.

[0196] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A probability density evaluation method for wind power, characterized in that: The method comprises: Obtaining output sample data of a wind farm and upper and lower limit data of wind power of the wind farm, and establishing a probability density function estimation model of wind power of the wind farm based on the logarithmic kernel differential homeomorphism estimation principle; According to the difference test, chi-square test and KS test between the probability model and the data observation distribution, a multi-objective optimization calculation model of the window width of the probability density function estimation model is established; Using a multi-objective particle swarm algorithm to solve the multi-objective optimization calculation model, and obtain the window width optimal solution set of the probability density function estimation model; Solving a final probability density function according to the window width optimal solution set and the probability density function estimation model, and evaluating the wind power of the wind farm according to the final probability density function; The step of obtaining the sample data of the wind farm output and the upper and lower limit data of the wind power of the wind farm, and establishing the probability density function estimation model of the wind farm specifically includes: Obtain the sample data P of the wind farm output w1 , P w2 ,…,P wn And the wind power upper and lower limit data [P min ,P max ]; According to the logarithmic kernel diffeomorphism estimation principle, the probability density estimation model of the wind power of the wind farm is calculated: Where P is the actual wind power output of the wind farm, P∈[P min ,P max ], f d,n (P) is the probability density function of the wind power of the wind farm, F(P) is the probability distribution function of the wind power of the wind farm, h d is the window width obtained after logarithmic diffeomorphic transformation of the output sample data of the wind farm, K() is the kernel function, P wi ∈[P min ,P max ], i = 1, 2, ..., n, n is the amount of output sample data, For the defined function, is a function The first derivative of .

2. The probability density evaluation method of wind power according to claim 1, characterized in that: The multi-objective optimization solution model is established based on the difference test, chi-square test and KS test between the probability model and the data observation distribution, specifically including: The output sample data P w1 , P w2 ,…,P wn Divide into m groups of non-intersecting data sets, and calculate the chi-square test statistic χ of the chi-square test 2 ; The sample data sample P w1 , P w2 ,…,P wn Sort from small to large and calculate the KS test statistic D of the KS test n ; The wind power upper and lower limit data [P min ,P max ] is divided into M intervals, and the output sample data P appearing in each interval is counted w1 , P w2 ,…,P wn The power histogram is used to calculate the difference test statistic MSR for the difference test between the probability model and the data observation distribution; Establish the multi-objective optimization solution model of the window width: objective function Objective function f2 = minD n , objective function f3 = minMSR and optimization variable h dx ; in, k i is the observed frequency of the ith set in m sets of non-intersecting data sets, p i is the theoretical probability value of the i-th set calculated by the probability density function, is a chi-square distribution with m-1 degrees of freedom. When the confidence level is given as α, for 1-α quantile of the output, n is the data volume of the output sample data; P w ' i The sample data P w1 , P w2 ,…,P wn The i-th sample after sorting from small to large; y i is the probability of the wind power of the wind farm in the ith interval calculated by the probability density function, x i is the probability of the wind power of the wind farm being in the i-th interval obtained by counting the power histogram; h dx ∈[h min ,h max ],h min 、h max are the upper and lower limits of the preset optimization variables respectively.

3. The probability density evaluation method of wind power according to claim 2, characterized in that: The method of solving the multi-objective optimization calculation model by using a multi-objective particle swarm algorithm to obtain the window width optimal solution set of the probability density function estimation model specifically includes: Establish a particle group, the particle group includes: a main particle group and three from the particle swarm Establish an external reserve set, initialize the particle group and the speed of each particle in the particle group, and solve the optimization variable h of the multi-objective optimization model according to the optimization variable h of the multi-objective optimization model. dx Each particle is randomly generated according to the constraints; According to the particle optimization update formula, the particles of the particle group are updated once, and the non-inferior extreme value particles obtained from the three slave particle groups are added to the main particle group, the three non-inferior particles from the slave particle group are added to the external reserve group, the non-inferior particles obtained from the main particle group are added to the external reserve set, the particle with the smallest value of the corresponding objective function f1 in the external reserve group is used as the global guiding particle of the slave particle group S1, the particle with the smallest value of the corresponding objective function f2 in the external reserve group is used as the global guiding particle of the slave particle group S2, the particle with the smallest value of the corresponding objective function f3 in the external reserve group is used as the global guiding particle of the slave particle group S3, and the particle with the highest sparsity in the external reserve group is used as the global guiding particle of the main particle group S0, completing an update optimization process; Repeat the particle group update optimization process until the number of iterations reaches a preset maximum value, and use the latest external reserve set as the window width optimal solution set of the probability density function estimation model; Where N' is the number of particles in the master particle group, N is the number of particles in each slave particle group, and the optimization update formula is: and and The particle swarm S at the k+1th and kth iterations are j The search speed of particle i, w j is the particle swarm S j The inertia weight, and The particle group S j The learning factor, r1 j and The particle group S j A random number that follows a uniform distribution on [0,1], and The particle group S j The current value of particle i at the kth iteration, the value of the individual local guiding particle, and the value of the global guiding particle.

4. The probability density evaluation method of wind power according to claim 2, characterized in that: Solving the final probability density function according to the window width optimal solution set and the probability density function estimation model, and evaluating the wind power of the wind farm according to the final probability density function, specifically includes: Selecting any solution in the window width optimal solution set as the final solution of the window width, substituting the optimal solution of the window width into the probability density function estimation model to obtain the final probability density function of the wind farm output; And the wind power of the wind farm is evaluated according to the final probability density function.

5. A wind power probability density assessment device, characterized in that: The device comprises: a density model module, an optimization model module, a solution module and an evaluation module; The density model module is used to obtain the output sample data of the wind farm and the upper and lower limit data of the wind power of the wind farm, and establish a probability density function estimation model of the wind power of the wind farm based on the logarithmic kernel differential homeomorphism estimation principle; The optimization model module is used to establish a multi-objective optimization calculation model for the window width of the probability density function estimation model according to the difference test, chi-square test and KS test between the probability model and the data observation distribution; The solution module is used to solve the multi-objective optimization calculation model using a multi-objective particle swarm algorithm to obtain a window width optimal solution set of the probability density function estimation model; The evaluation module is used to solve the final probability density function according to the window width optimal solution set and the probability density function estimation model, and evaluate the wind power of the wind farm according to the final probability density function; the density model module is specifically used to: Obtain the sample data P of the wind farm output w1 , P w2 ,…,P wn And the wind power upper and lower limit data [P min ,P max ]; According to the logarithmic kernel diffeomorphism estimation principle, the probability density estimation model of the wind power of the wind farm is calculated: Where P is the actual wind power output of the wind farm, P∈[P min ,P max ], f d,n (P) is the probability density function of the wind power of the wind farm, F(P) is the probability distribution function of the wind power of the wind farm, h d is the window width obtained after logarithmic diffeomorphic transformation of the output sample data of the wind farm, K() is the kernel function, P wi ∈[P min ,P max ], i = 1, 2, ..., n, n is the amount of output sample data, For the defined function, is a function The first derivative of .

6. The wind power probability density evaluation device according to claim 5, characterized in that: The optimization model module is specifically used for: The output sample data P w1 , P w2 ,…,P wn Divide into m groups of non-intersecting data sets, and calculate the chi-square test statistic χ of the chi-square test 2 ; The sample data sample P w1 , P w2 ,…,P wn Sort from small to large and calculate the KS test statistic D of the KS test n ; The wind power upper and lower limit data [P min ,P max ] is divided into M intervals, and the output sample data P appearing in each interval is counted w1 , P w2 ,…,P wn The power histogram is used to calculate the difference test statistic MSR for the difference test between the probability model and the data observation distribution; Establish the multi-objective optimization solution model of the window width: objective function Objective function f2 = minD n , objective function f3 = minMSR and optimization variable h dx ; in, k i is the observed frequency of the ith set in m sets of non-intersecting data sets, p i is the theoretical probability value of the i-th set calculated by the probability density function, is a chi-square distribution with m-1 degrees of freedom. When the confidence level is given as α, for 1-α quantile of the output, n is the data volume of the output sample data; P w ' i The sample data P w1 , P w2 ,…,P wn The i-th sample after sorting from small to large; y i is the probability of the wind power of the wind farm in the ith interval calculated by the probability density function, x i is the probability of the wind power of the wind farm being in the i-th interval obtained by counting the power histogram; h dx ∈[h min ,h max ],h min 、h max are the upper and lower limits of the preset optimization variables respectively.

7. A terminal device, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor implements the probability density assessment method of wind power as claimed in any one of claims 1 to 4 when executing the computer program.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the probability density assessment method for wind power according to any one of claims 1 to 4.

Images