Network following controller parameter optimization method and device
By adopting a method based on the Liyapunov second method and the sum of squares theory in the power system, combining axial search and pattern search to optimize the network controller parameters, the problem of conservative attraction domain estimation results in the power system stability assessment is solved. Improves the transient stability and computing efficiency of the system.
Patent Information
- Application Number
- CN202411005449.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-25
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2044-07-25
AI Technical Summary
The existing power system stability evaluation method is too conservative in the attraction domain estimation after failure, and is greatly affected by the system vector field conditions, resulting in a reduced system operation efficiency.
The extended intraditional method and square sum theory based on the Liyapunov second method are adopted, combined with axial search and pattern search, and the following network controller parameters are optimized to improve the accuracy and efficiency of attractive domain estimation.
By optimizing the parameters of the network controller, expanding the attraction domain, improving the dynamic characteristics uniformity of the power system, and enhancing the system's transient stability and computing efficiency.
Smart Images

Figure CN118983774A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of power system automation, and in particular to a method and device for optimizing parameters of a grid-following controller. Background Art
[0002] Transient stability is one of the important issues to be solved in power systems. Analyzing the stability of power systems helps ensure that power systems operate safely and efficiently. Without analyzing stability, the impact of faults on stability cannot be estimated, and the system may become unstable under unexpected circumstances. In addition, if system stability is ensured without analyzing system stability, the operation of the system must be conservative, thereby reducing efficiency. With the rapid development of social economy and science and technology, the scale of power systems is expanding, the proportion of renewable energy is gradually increasing, and the number of power electronic equipment in the power grid is increasing, making the stability problem of power systems more and more complicated.
[0003] Existing stability assessment methods can be divided into non-Lyapunov function-based methods and Lyapunov function-based methods. The Lyapunov function-based method is committed to finding a stable region of the equilibrium point after the fault. If the initial state of the system after the fault is within this stable region, it can be asserted that the system will eventually converge to the equilibrium point. The sum-of-squares-based method is an emerging algorithm tool proposed in the past decade or so. It can systematically construct the Lyapunov function and determine a sufficiently large approximate attraction domain in the power system model considering the transfer conductance. However, the result of the attraction domain estimation is greatly affected by the state of the system vector field. The equilibrium surface or unstable equilibrium point close to the equilibrium point in the vector field may hinder the expansion of the attraction domain to be estimated, resulting in an overly conservative result of the attraction domain estimation. Summary of the invention
[0004] In order to solve the above technical problems, an embodiment of the present invention provides a method and device for optimizing parameters of a grid controller, which improves the attraction domain estimation effect of the grid controller on the power system by optimizing the parameters of the grid controller in the power grid system.
[0005] A first aspect of an embodiment of the present invention provides a method for optimizing network controller parameters, the method comprising:
[0006] Construct a dynamic model of the grid controller in the power system to be analyzed, perform dynamic analysis on the dynamic model of the grid controller, and obtain the state vector equation;
[0007] The extended inner domain method based on Lyapunov's second method is used in combination with the sum of squares theory to solve the attraction domain of the power system to be analyzed, and the estimation result of the attraction domain in the power system to be analyzed is obtained;
[0008] The time domain simulation of the system fault state of the power system to be analyzed is performed to obtain the true critical removal time and the state variable values at each moment, and the state variable values at each moment are substituted into the Lyapunov function to obtain multiple Lyapunov function values. The moment corresponding to the Lyapunov function value that meets the preset conditions is selected as the estimated critical removal time, and the error index of the attraction domain estimation result is obtained using the true critical removal time and the estimated critical removal time.
[0009] If the error index is greater than the preset index, the optimal parameter search is performed based on the state vector equation using axial search. When the step size is less than the first preset value, the initial Lyapunov function obtained last time is used to solve the attraction domain of the power system to be analyzed, and an updated attraction domain estimation result is obtained. Based on the updated attraction domain estimation result, the pattern search and axial search are used to continue searching for the optimal solution of the parameters of the grid controller until the step size is less than the second preset value, and the optimal parameter result is obtained.
[0010] In a possible implementation manner of the first aspect, a dynamic analysis is performed on a dynamic model of the network tracking controller to obtain a state vector equation, including:
[0011] The dynamic model of the grid controller is dynamically analyzed, and the state vector equations of the grid control power supply are obtained in the dynamic analysis. The state vector equations are:
[0012]
[0013] Among them, Δδ c is the controller equivalent relative phase angle, Δδ c =θ pll -θ ref , is the equivalent speed deviation of the controller, is the speed deviation of the system reference synchronous generator, is the q-axis voltage at the common coupling point where the grid-following control power source is connected to the grid, f n is the system rated frequency, is the rated speed of the system, K p and K i It is an adjustable parameter of the network controller.
[0014] The state vector equation of the reference synchronous generator is:
[0015]
[0016] in, is the relative speed of the reference synchronous generator, H ref is the reference synchronous generator moment of inertia, is the reference synchronous generator damping coefficient, is the reference synchronous generator mechanical power input, is the reference synchronous generator electrical power output.
[0017] The state vector equations of the remaining synchronous generators in the system are:
[0018]
[0019] Among them, δ i is the phase angle of the synchronous generator, is the relative speed of the synchronous generator, H i is the moment of inertia of the synchronous generator, is the damping coefficient of the synchronous generator, is the synchronous generator mechanical power input, The electrical power output of the synchronous generator.
[0020] In a possible implementation of the first aspect, an extended inner domain method based on Lyapunov's second method is used in combination with the sum of squares theory to solve the attraction domain of the power system to be analyzed, and an estimation result of the attraction domain in the power system to be analyzed is obtained, including:
[0021] Set the criterion for the end of the loop γ 0 :=0,i:=1,j:=1, select the initial Lyapunov function, and use the attraction domain estimation problem model to solve it, where the attraction domain estimation problem model is:
[0022] find(V 0 ,λ1,λ3,s2)
[0023] st
[0024]
[0025] in, is the set of square sum polynomials, is a set of n-1 dimensional polynomial vectors, G is the equality constraint generated in the process of detrigonometric function of the system model, β is a positive number, r(z) is a positive definite polynomial, λ1 and λ3 are n-1 dimensional polynomial variables added to the equality constraint G in the positive point theorem, and s2 is the square sum polynomial variable added to the inequality constraint β-r(z) in the positive point theorem; and is a square sum polynomial,
[0026] set up Setting V (i) (z):=V (i-1) (z), γ: = γ(i-1) , transform the attraction domain estimation problem into the first square sum planning problem, obtain the first square sum optimization problem model and solve it to obtain the first solution result, where the first square sum optimization problem is:
[0027] (SOSP1)max c
[0028] st
[0029]
[0030] in, is the set of square sum polynomials, is a set of n-1 dimensional polynomial vectors, G is the equality constraint generated in the process of detrigonometric function of the system model, β and c are positive numbers, p is a positive definite polynomial, λ2 and λ3 are n-1 dimensional polynomial variables added to the equality constraint G in the positive point theorem, s1, s2, s3 are square sum polynomial variables added to the inequality constraint in the positive point theorem, is a square sum polynomial, which is used to ensure that each constraint is semi-positive definite.
[0031] Taking the first solution result as a variable, the first square sum planning problem is transformed into the second square sum planning problem, and the second square sum optimization problem model is obtained and solved to obtain the second solution result, wherein the second square sum optimization problem model is:
[0032] (SOSP2)maxβ
[0033] st
[0034]
[0035] in, is the set of square sum polynomials, is a set of n-1 dimensional polynomial vectors, G is the equality constraint generated in the process of detrigonometric function of the system model, β and c are positive numbers, p is a positive definite polynomial, λ1, λ2, λ3 are n-1 dimensional polynomial variables added to the equality constraint G in the positive point theorem, s1, s2, s3 are square sum polynomial variables added to the inequality constraint in the positive point theorem, and is a square sum polynomial, which is used to ensure that each constraint is semi-positive definite.
[0036] The second solution results β and V(z) are stored as β (i) and V (i) (z), if β (i) -β (i-1) <∈β , then if j>1 and and If the maximum absolute value of the coefficient is less than the preset criterion, the estimation result of the attraction domain and the current Lyapunov function are obtained. If not, V is set. (0) (z):=V (i) (z), P 0 (z):=V (i )(z),β (0) :=c (i ), i: = 1, j = j + 1, re-solving the first square sum optimization problem model and the second square sum optimization problem model until the attraction domain estimation result and the Lyapunov function are obtained;
[0037] If β (i) -β (i-1) >∈ β , otherwise set i=i+1, and re-solve the first sum of squares optimization problem model and the second sum of squares optimization problem model until the attraction domain estimation result and Lyapunov function are obtained.
[0038] In a possible implementation manner of the first aspect, a calculation formula for the error index is:
[0039]
[0040] Among them, t ref represents the real critical removal time, t cal Represents the estimated critical resection time.
[0041] In a possible implementation manner of the first aspect, performing optimal parameter search by using axial search includes:
[0042] Along K p Direction search, if Then like Then Otherwise,
[0043] Along K i Direction search. Then like Then Otherwise, Among them, K p and K i is an adjustable parameter of the network controller, and d is the step size of the axial search.
[0044] In a possible implementation manner of the first aspect, searching for an optimal solution for parameters of a network controller by using a pattern search includes:
[0045] set up Along the y k =x k -x k-1 Direction search, if Δ(x k -αy k )<Δ(y k ), let x k+1 =x k -αy k Otherwise, let x k+1 =x k .
[0046] A second aspect of an embodiment of the present invention provides a network controller parameter optimization device, the device comprising:
[0047] A construction module is used to construct a dynamic model of a grid controller in the power system to be analyzed, and to dynamically analyze the dynamic model of the grid controller to obtain a state vector equation;
[0048] An estimation module, used for solving the attraction domain of the power system to be analyzed by using an extended inner domain method based on Lyapunov's second method combined with the sum of squares theory, and obtaining an estimation result of the attraction domain in the power system to be analyzed;
[0049] The error determination module is used to perform time domain simulation of the system fault state of the power system to be analyzed, obtain the real critical removal time and the state variable values at each moment, substitute the state variable values at each moment into the Lyapunov function, obtain multiple Lyapunov function values, select the moment corresponding to the Lyapunov function value that meets the preset conditions as the estimated critical removal time, and use the real critical removal time and the estimated critical removal time to obtain the error index of the attraction domain estimation result;
[0050] The parameter calculation module is used to search for the optimal parameters based on the state vector equation using axial search if the error index is greater than the preset index. When the step size is less than the first preset value, the initial Lyapunov function obtained last time is used to solve the attraction domain of the power system to be analyzed to obtain an updated attraction domain estimation result. Based on the updated attraction domain estimation result, the optimal solution of the parameters of the grid controller is continued to be searched using pattern search and axial search until the step size is less than the second preset value to obtain the optimal parameter result.
[0051] In a possible implementation manner of the second aspect, a dynamic analysis is performed on a dynamic model of the network controller to obtain a state vector equation, including:
[0052] The dynamic model of the grid controller is dynamically analyzed, and the state vector equations of the grid control power supply are obtained in the dynamic analysis. The state vector equations are:
[0053]
[0054] Among them, Δδ c is the controller equivalent relative phase angle, Δδ c =θ pll -θ ref , is the equivalent speed deviation of the controller, is the speed deviation of the system reference synchronous generator, is the q-axis voltage at the common coupling point where the grid-following control power source is connected to the grid, f n is the system rated frequency, is the rated speed of the system, K p and K i It is an adjustable parameter of the network controller.
[0055] The state vector equation of the reference synchronous generator is:
[0056]
[0057] in, is the relative speed of the reference synchronous generator, H ref is the reference synchronous generator moment of inertia, is the reference synchronous generator damping coefficient, is the reference synchronous generator mechanical power input, is the reference synchronous generator electrical power output.
[0058] The state vector equations of the remaining synchronous generators in the system are:
[0059]
[0060] Among them, δ i is the phase angle of the synchronous generator, is the relative speed of the synchronous generator, H i is the moment of inertia of the synchronous generator, is the damping coefficient of the synchronous generator, is the synchronous generator mechanical power input, The electrical power output of the synchronous generator.
[0061] A third aspect of an embodiment of the present invention provides a computer device, including:
[0062] Memory for storing computer programs;
[0063] The processor is used to implement the network controller parameter optimization method of the first aspect when executing the computer program.
[0064] A fourth aspect of an embodiment of the present invention provides a storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the network controller parameter optimization method according to the first aspect are implemented.
[0065] The embodiment of the present invention provides a method of constructing a dynamic model of a grid controller in an electric power system to be analyzed, dynamically analyzing the dynamic model of the grid controller to obtain a state vector equation, adopting an extended inner domain method based on Lyapunov's second method, and combining the square sum theory to solve the attraction domain of the electric power system to be analyzed, and obtaining an estimation result of the attraction domain in the electric power system to be analyzed; performing a time domain simulation of the system fault state of the electric power system to be analyzed to obtain the true critical resection time and the state variable values at each moment, substituting the state variable values at each moment into the Lyapunov function to obtain multiple Lyapunov function values, selecting the moment corresponding to the Lyapunov function value that meets the preset conditions as the estimated critical resection time, and using the true critical resection time to estimate the critical resection time. The error index of the attraction domain estimation result is obtained by dividing the time and estimating the critical cut-off time; if the error index is greater than the preset index, the optimal solution of the parameters of the grid controller is searched by axial search based on the state vector equation. When the step size is less than the first preset value, the initial Lyapunov function obtained last time is used to solve the attraction domain of the power system to be analyzed, and an updated attraction domain estimation result is obtained. Based on the updated attraction domain estimation result, the optimal solution of the parameters of the grid controller is continued to be searched by pattern search and axial search until the step size is less than the second preset value to obtain the optimal parameter result. This method improves the attraction domain estimation effect of the grid controller on the power system by optimizing the parameters of the grid controller in the power grid system. BRIEF DESCRIPTION OF THE DRAWINGS
[0066] Figure 1 : A flow chart of a parameter optimization method according to an embodiment of a network controller parameter optimization method provided by the present invention;
[0067] Figure 2 : A structural block diagram of a network controller according to an embodiment of a network controller parameter optimization method provided by the present invention;
[0068] Figure 3 : A flow chart of an algorithm for estimating the power system attraction domain of a grid controller according to an embodiment of a method for optimizing grid controller parameters provided by the present invention;
[0069] Figure 4 : A flow chart of a network controller parameter optimization algorithm for a network controller of an embodiment of a network controller parameter optimization method provided by the present invention;
[0070] Figure 5 : A calculation example data diagram of an embodiment of the network controller parameter optimization method provided by the present invention;
[0071] Figure 6 : A diagram of an example of an example of an embodiment of the method for optimizing network controller parameters provided by the present invention;
[0072] Figure 7 : A device block diagram of an embodiment of the network controller parameter optimization method provided by the present invention. DETAILED DESCRIPTION
[0073] 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.
[0074] At present, although the method based on the sum of squares can systematically construct the Lyapunov function and determine a sufficiently large approximate attraction domain in the power system model considering the transfer conductance, the result of the attraction domain estimation is greatly affected by the state of the system vector field. The equilibrium surface or unstable equilibrium point close to the equilibrium point in the vector field may hinder the expansion of the estimated attraction domain, resulting in an overly conservative result of the attraction domain estimation.
[0075] Based on this, an embodiment of the present application provides a method and device for optimizing the parameters of a network controller, which aims to find suitable network controller parameters when running an algorithm for estimating the attraction domain so that the dynamic characteristics of the entire system are more uniform, so as to obtain a maximized attraction domain estimation result and improve computational efficiency.
[0076] Please refer to Figure 1 , which is a flow chart of an embodiment of a method for optimizing network controller parameters provided by an embodiment of the present invention, including steps S101 to S104, each of which is specifically as follows:
[0077] S101, constructing a dynamic model of a grid controller in the power system to be analyzed, and dynamically analyzing the dynamic model of the grid controller to obtain a state vector equation.
[0078] In this embodiment, a dynamic model of the network controller in the system to be analyzed is designed, and a transfer function and a state vector equation group of the network controller are derived according to the dynamic model.
[0079] In some embodiments, step S101 of "dynamically analyzing the dynamic model of the network tracking controller to obtain a state vector equation" includes but is not limited to the following steps:
[0080] The dynamic model of the grid controller is dynamically analyzed, and the state vector equations of the grid control power supply are obtained in the dynamic analysis. The state vector equations are:
[0081]
[0082] Among them, Δδ c is the controller equivalent relative phase angle, Δδ c =θ pll -θ ref , is the equivalent speed deviation of the controller, is the speed deviation of the system reference synchronous generator, is the q-axis voltage at the common coupling point where the grid-following control power source is connected to the grid, f n is the system rated frequency, is the rated speed of the system, K p and K i It is an adjustable parameter of the network controller.
[0083] The state vector equation of the reference synchronous generator is:
[0084]
[0085] in, is the relative speed of the reference synchronous generator, H ref is the reference synchronous generator moment of inertia, is the reference synchronous generator damping coefficient, is the reference synchronous generator mechanical power input, is the reference synchronous generator electrical power output.
[0086] The state vector equations of the remaining synchronous generators in the system are:
[0087]
[0088] Among them, δ i is the phase angle of the synchronous generator, is the relative speed of the synchronous generator, H i is the moment of inertia of the synchronous generator, is the damping coefficient of the synchronous generator, is the synchronous generator mechanical power input, The electrical power output of the synchronous generator.
[0089] Specifically, the model structure of the network controller is as follows: Figure 2 As shown, in Figure 2 middle, is the a-phase component of the common connection point voltage, is the b-phase component of the common connection point voltage, is the c-phase component of the common connection point voltage, is the d-axis component of the common connection point voltage, is the q-axis component of the voltage at the common connection point, K p is the proportional adjustment coefficient of the grid controller, K i is the integral adjustment coefficient of the grid-following controller, ω n is the rated speed, f n is the rated frequency, θ pll is the phase difference of the phase-locked loop.
[0090] The dynamic characteristics of the system include the speed and phase difference of the synchronous generator and the grid-following control power supply. The state vector equation of the power supply connected to the power system using the grid-following controller in the dynamic analysis is:
[0091]
[0092] Among them, Δδ c is the controller equivalent relative phase angle, Δδ c =θ pll -θ ref , is the equivalent speed deviation of the controller, is the speed deviation of the system reference synchronous generator, is the q-axis voltage at the common coupling point where the grid-following control power source is connected to the grid, f n is the system rated frequency, is the rated speed of the system, K p and K i It is an adjustable parameter of the network controller.
[0093] At the same time, the models of the system reference synchronous generator and the remaining synchronous generators are constructed. The state vector equation of the system reference synchronous generator is:
[0094]
[0095] in, is the relative speed of the reference synchronous generator, H ref is the reference synchronous generator moment of inertia, is the reference synchronous generator damping coefficient, is the reference synchronous generator mechanical power input, is the reference synchronous generator electrical power output.
[0096] The state vector equations of the remaining synchronous generators in the system are:
[0097]
[0098] Among them, δ i is the phase angle of the synchronous generator, is the relative speed of the synchronous generator, H i is the moment of inertia of the synchronous generator, is the damping coefficient of the synchronous generator, is the synchronous generator mechanical power input, The electrical power output of the synchronous generator.
[0099] S102. Using the extended inner domain method based on Lyapunov's second method and combining it with the sum of squares theory, the attraction domain of the power system to be analyzed is solved to obtain an estimation result of the attraction domain in the power system to be analyzed.
[0100] In this embodiment, an algorithm for solving the controller parameter optimization index is designed, and the system attraction domain estimation algorithm adopts an extended inner domain method based on Lyapunov's second method, combined with the square sum theory to solve the attraction domain of the power system to be analyzed, and obtains the attraction domain estimation result in the power system to be analyzed.
[0101] In some embodiments, in step S102, "using the extended inner domain method based on Lyapunov's second method and combining the square sum theory to solve the attraction domain of the power system to be analyzed, and obtaining the attraction domain estimation result in the power system to be analyzed" includes but is not limited to the following steps:
[0102] Set the criterion for the end of the loop γ 0 :=0,i:=1,j:=1, select the initial Lyapunov function, and use the attraction domain estimation problem model to solve it, where the attraction domain estimation problem model is:
[0103] find(V 0 ,λ1,λ3,s2)
[0104] st
[0105]
[0106] in, is the set of square sum polynomials, is a set of n-1 dimensional polynomial vectors, G is the equality constraint generated in the process of detrigonometric function of the system model, β is a positive number, r(z) is a positive definite polynomial, λ1 and λ3 are n-1 dimensional polynomial variables added to the equality constraint G in the positive point theorem, and s2 is the square sum polynomial variable added to the inequality constraint β-r(z) in the positive point theorem; and is a square sum polynomial,
[0107] set up Setting V (i) (z):=V (i-1) (z), γ: = γ (i-1), transform the attraction domain estimation problem into a first square sum planning problem, obtain a first square sum optimization problem model and solve it to obtain a first solution result, wherein the first square sum optimization problem is:
[0108] (SOSP1)max c
[0109] st
[0110]
[0111] in, is the set of square sum polynomials, is a set of n-1 dimensional polynomial vectors, G is the equality constraint generated in the process of detrigonometric function of the system model, β and c are positive numbers, p is a positive definite polynomial, λ2 and λ3 are n-1 dimensional polynomial variables added to the equality constraint G in the positive point theorem, s1, s2, s3 are square sum polynomial variables added to the inequality constraint in the positive point theorem, is a square sum polynomial, which is used to ensure that each constraint is semi-positive definite.
[0112] Taking the first solution result as a variable, the first square sum planning problem is transformed into a second square sum planning problem, and a second square sum optimization problem model is obtained and solved to obtain a second solution result, wherein the second square sum optimization problem model is:
[0113] (SOSP2)maxβ
[0114] st
[0115]
[0116] in, is the set of square sum polynomials, is a set of n-1 dimensional polynomial vectors, G is the equality constraint generated in the process of detrigonometric function of the system model, β and c are positive numbers, p is a positive definite polynomial, λ1, λ2, λ3 are n-1 dimensional polynomial variables added to the equality constraint G in the positive point theorem, s1, s2, s3 are square sum polynomial variables added to the inequality constraint in the positive point theorem, and is a square sum polynomial, which is used to ensure that each constraint is semi-positive definite.
[0117] The second solution results β and V(z) are stored as β (i) and V (i) (z), if β (i) -β (i-1) <∈β , then if j>1 and and If the maximum absolute value of the coefficient is less than the preset criterion, the estimation result of the attraction domain and the current Lyapunov function are obtained. If not, V is set. (0) (z):=V (i) (z), P 0 (z):=V (i) (z), β (0) :=c (i) , i:=1, j=j+1, resolving the first sum-of-squares optimization problem model and the second sum-of-squares optimization problem model until the attraction domain estimation result and the Lyapunov function are obtained;
[0118] If β (i) -β (i-1) >∈ β , otherwise set i=i+1, and re-solve the first sum of squares optimization problem model and the second sum of squares optimization problem model until the attraction domain estimation result and the Lyapunov function are obtained.
[0119] Specifically, the process of the power system attraction region estimation algorithm is as follows: Figure 3 As shown, Feasratio is the quality index of SOSP2 optimization problem solution. The algorithm adopts the extended inner domain method based on Lyapunov's second method and combines the sum of squares theory for optimization solution. The basic iterative process of the algorithm is as follows:
[0120] Step 1: Set the criteria for loop end And set γ 0 :=0,i:=1,j:=1. The initial Lyapunov function V is selected by solving the following square sum optimization problem 0 .
[0121] find(V 0 ,λ1,λ3,s2)
[0122] st
[0123]
[0124] in, is the set of square sum polynomials, It is a set of n-1 dimensional polynomial vectors. The algorithm converts the original multiple equality constraints and inequality constraints into polynomial form through the P-satz theorem. G is the equality constraint generated in the process of de-trigonometric function of the system model. β is a positive number, r(z) is a positive definite polynomial, λ1 and λ3 are n-1 dimensional polynomial variables added to the equality constraint G in the positive point theorem, and s2 is the square sum polynomial variable added to the inequality constraint β-r(z) in the positive point theorem; and is a square sum polynomial, which is used to ensure that each constraint is semi-positive definite.
[0125] Step 2: Setup
[0126] Step 3: Set V (i) (z):=V (i-1) (z), γ: = γ (i-1) , and solve the following sum-of-squares optimization problem
[0127] (SOSP1)max c
[0128] st
[0129]
[0130] in, is the set of square sum polynomials, is a set of n-1 dimensional polynomial vectors, G is the equality constraint generated in the process of detrigonometric function of the system model, β and c are positive numbers, p is a positive definite polynomial, λ2 and λ3 are n-1 dimensional polynomial variables added to the equality constraint G in the positive point theorem, s1, s2, s3 are square sum polynomial variables added to the inequality constraint in the positive point theorem, is a square sum polynomial, which is used to ensure that each constraint is semi-positive definite.
[0131] Store the solution results c, s2(z), and s3(z) as c (i) ,
[0132] Step 4: Set c: = c (i) , And solve the following sum-of-squares optimization problem:
[0133] (SOSP2)maxβ
[0134] st
[0135]
[0136] in, is the set of square sum polynomials, is a set of n-1 dimensional polynomial vectors, G is the equality constraint generated in the process of detrigonometric function of the system model, β and c are positive numbers, p is a positive definite polynomial, λ1, λ2, λ3 are n-1 dimensional polynomial variables added to the equality constraint G in the positive point theorem, s1, s2, s3 are square sum polynomial variables added to the inequality constraint in the positive point theorem, and is a square sum polynomial, which is used to ensure that each constraint is semi-positive definite.
[0137] in is the set of square sum polynomials, is a set of n-1 dimensional polynomial vectors, G is the equality constraint generated in the process of detrigonometric function of the system model, β and c are positive numbers, p is a positive definite polynomial, λ1, λ2, λ3 are n-1 dimensional polynomial variables added to the equality constraint G in the positive point theorem, s1, s2, s3 are square sum polynomial variables added to the inequality constraint in the positive point theorem, and is a square sum polynomial, which is used to ensure that each constraint is semi-positive definite.
[0138] The solution results β and V(z) are stored as β (i) and V (i) (z).
[0139] Step 5: If β (i) -β (i-1) <∈ β , then go to step 6, otherwise set i=i+1 and go to step 3.
[0140] Step 6: If j>1 and and The maximum absolute value of the coefficient is less than ∈ l , go to step 7. Otherwise, set V (0) (z):=V (i) (z), P 0 (z):=V (i) (z), β (0) :=c (i) , i:=1, j=j+1, and go to step 2.
[0141] Step 7: Output V (i) (z) and c (i) .
[0142] S103. Perform time domain simulation of the fault state of the power system to be analyzed to obtain the true critical resection time and the state variable values at each moment, substitute the state variable values at each moment into the Lyapunov function to obtain multiple Lyapunov function values, select the moment corresponding to the Lyapunov function value that meets the preset conditions as the estimated critical resection time, and use the true critical resection time and the estimated critical resection time to obtain the error index of the attraction domain estimation result.
[0143] In this embodiment, the parameter optimization index of the grid controller is the error of the critical clearing time (CCT) of a specific fault in the power system.
[0144] After obtaining the attraction domain estimation result using the above power system attraction domain estimation algorithm, a specific fault point and fault type are selected in the power system, and the post-fault state of the power system is simulated in the time domain to obtain its accurate critical removal time. At the same time, the state variable values at each moment in the time domain simulation are substituted into the Lyapunov function obtained by the attraction domain estimation, and the moment when the Lyapunov function value exceeds 1 for the first time is selected as the estimated critical removal time. The error index of the attraction domain estimation result is obtained by using the real critical removal time and the estimated critical removal time.
[0145] In some embodiments, the "error indicator" in step S103 further includes but is not limited to the following steps:
[0146] The calculation formula of the error index is:
[0147]
[0148] Among them, t ref represents the real critical removal time, t cal Represents the estimated critical resection time.
[0149] Specifically, the calculation formula of the error index is:
[0150]
[0151] Among them, t ref represents the real critical removal time, t cal Represents the estimated critical resection time.
[0152] S104. If the error index is greater than the preset index, the optimal solution of the parameters of the grid controller is searched by axial search based on the state vector equation. When the step size is less than the first preset value, the initial Lyapunov function obtained last time is used to solve the attraction domain of the power system to be analyzed, and an updated attraction domain estimation result is obtained. Based on the updated attraction domain estimation result, the optimal solution of the parameters of the grid controller is continued to be searched by pattern search and axial search until the step size is less than the second preset value, and the optimal parameter result is obtained.
[0153] In this embodiment, when the error of the estimation result of the attraction domain is large, the pattern search method is used to search for the optimal solution of the two parameters of the network controller. Specifically, the algorithm flow of the network controller parameter optimization is as follows: Figure 4 As shown in the figure, the pattern search method is divided into two parts, axial search and pattern search, which run alternately. Each cycle includes one axial search and one pattern search, so as to approach the optimal value faster. At the same time, the internal iteration process of the attraction domain estimation algorithm is modified during the axial search iteration. When the search step size is reduced to a certain extent, the initial Lyapunov function obtained in the previous iteration is used as the initial Lyapunov function of the current iteration, thereby reducing the calculation time. That is, when the step size d of the axial search is reduced to a certain extent, the iterative process of the attraction domain estimation algorithm skips steps 1 and 2, and uses the initial Lyapunov function V calculated in the previous step of the axial search instead. 0 and l1 and l2 are input into step 3.
[0154] The above-mentioned network controller parameter optimization algorithm selects appropriate parameter optimization indicators, so that the system attraction domain solution and accuracy judgment can be automatically calculated, so that the optimization algorithm can be automatically calculated, reducing the risk of human error and improving the efficiency of the algorithm. Secondly, the modification of the attraction domain estimation algorithm can effectively alleviate the problem of the algorithm facing the excessive calculation burden of higher-order systems, so that the system can quickly obtain the optimal controller parameters.
[0155] In some embodiments, the step S104 of "using axial search to search for optimal parameters" further includes but is not limited to the following steps:
[0156] Along K p Direction search, if Then like Then Otherwise,
[0157] Along K i Direction search. Then like Then Otherwise, Among them, K p and K i is an adjustable parameter of the network controller, and d is the step size of the axial search.
[0158] Specifically, when the algorithm is running, the network controller parameter K is first set p and K i The initial value of , the axial search strategy is:
[0159] Step 11: Along the K p Direction search, if Then like Then Otherwise,
[0160] Step 12: Along the K i Direction search. Then like Then Otherwise, Among them, K p and K i is an adjustable parameter of the network controller, and d is the step size of the axial search.
[0161] In some embodiments, in step S104, "searching for the optimal solution of the parameters of the network controller by using a pattern search" further includes but is not limited to the following steps:
[0162] set up Along the y k =x k -x k-1 Direction search, if Δ(x k -αy k )<Δ(y k ), let x k+1 =x k -αy k Otherwise, let x k+1 =x k .
[0163] Specifically, the pattern search strategy is:
[0164] set up Along the y k =x k -x k-1 Direction search, if Δ(x k -αy k )<Δ(y k ), let x k+1 =x k -αy k Otherwise, let x k+1 =x k .
[0165] The parameters of the grid-following controller include integral coefficient and proportional coefficient. The parameterization index is selected as the error estimated by the attraction domain of the critical removal time of a specified fault at a specified node in the power grid.
[0166] This method designs the dynamic model of the grid controller, derives the controller transfer function and state vector equations, and then designs the algorithm for solving the controller parameter optimization index. Taking the error of the system attraction domain estimation result as an indicator, select the representative nodes (or multiple nodes) and fault types in the system, calculate their critical cut-off time (CCT) through the attraction domain obtained by time domain simulation and algorithm estimation, and use the error between the two as the indicator of the error of the attraction domain estimation result. Finally, design the optimization adjustment strategy of the grid controller parameters. Set the initial value of the controller parameters and run the attraction domain estimation algorithm, and evaluate the accuracy of the attraction domain. Use the pattern search method to search for the optimal solution of the controller parameters; at the same time, when the adjustment range of the pattern search method is reduced to the set accuracy, modify the internal iteration process of the attraction domain estimation algorithm to reduce the calculation time. At the same time, when running the algorithm for estimating the attraction domain, by finding the appropriate grid controller parameters, the dynamic characteristics of the whole system are more uniform, so as to obtain the maximum attraction domain estimation result, improve the calculation efficiency, and improve the attraction domain estimation effect of the grid controller on the power system.
[0167] The present application also implements the technical solution provided in the embodiment of the present application, selects a 3-machine 9-node system model, in which there are two synchronous generators and a motor controlled by the grid, and sets the initial parameters to Step size d=6, ε1=1, ε2=1. Figure 5 is the parameter K in the iterative process of the optimization algorithm p ,K i The selected values, among which parameter groups 2, 3, 4 and parameter groups 6, 7, 8, 9, 10, 11 are the results of axial search, and parameter group 5 is the result of axial search. During the iteration process, the optimization index Δ gradually decreases, and the step size also gradually decreases as the search fails, and finally the optimal result is obtained.
[0168] Select the parameter groups 1, 2, 4, 5, 10, and 11 with the reduced optimization index Δ to draw the attraction domain image as shown in Figure 6 As shown in the figure, the horizontal and vertical coordinates are selected respectively. As the optimization index Δ decreases, the range of the attraction domain gradually expands. Only when the adjustment range from data group 10 to data group 11 is extremely small, there is a non-positive correlation, which is enough to prove the rationality of the selection of the optimization index Δ.
[0169] This method can give full play to the maximum performance of the doubly fed wind turbine converter. In the early stage of transient oscillation of the power system, the four output variables of the doubly fed wind turbine can be converged to a certain neighborhood near the equilibrium point at the fastest speed through the switch controller. Then switch to a conventional vector controller to give full play to the optimality of the conventional vector controller near the equilibrium point, so that the system asymptotically stabilizes to the original equilibrium point. In addition, the switch controller proposed by this method only contains logical operations, so that the phase lag between its output and input is smaller than that of the conventional vector controller. Therefore, the switch controller can respond to the oscillation of the output variables of the doubly fed wind turbine more quickly. At the same time, the control signal of the switch controller has only two values, which facilitates the transmission of the control signal.
[0170] The application of the above method in the coordinated control of doubly-fed wind turbines can greatly improve the transient stability of the power system operation containing large-scale wind power.
[0171] It should be understood that although Figure 1 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 1 At least part of the steps may include multiple steps or multiple stages. These steps or stages are not necessarily performed at the same time, but can be performed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed in turn or alternately with other steps or at least part of the steps or stages in other steps.
[0172] In some embodiments, Figure 7 As shown, it shows a block diagram of a network controller parameter optimization device 700 provided in an embodiment of the present application, including: a construction module 701, an estimation module 702, an error determination module 703 and a parameter calculation module 704, wherein:
[0173] A construction module 701 is used to construct a dynamic model of a grid controller in the power system to be analyzed, and to dynamically analyze the dynamic model of the grid controller to obtain a state vector equation;
[0174] An estimation module 702 is used to adopt an extended inner domain method based on Lyapunov's second method and combine it with the square sum theory to solve the attraction domain of the power system to be analyzed, and obtain an estimation result of the attraction domain in the power system to be analyzed;
[0175] The error determination module 703 is used to perform a time domain simulation of the system fault state of the power system to be analyzed, obtain the real critical removal time and the state variable values at each time, substitute the state variable values at each time into the Lyapunov function, obtain multiple Lyapunov function values, select the time corresponding to the Lyapunov function value that meets the preset conditions as the estimated critical removal time, and use the real critical removal time and the estimated critical removal time to obtain the error index of the attraction domain estimation result;
[0176] The parameter calculation module 704 is used to search for the optimal solution of the parameters of the grid controller based on the state vector equation using axial search if the error index is greater than the preset index. When the step size is less than the first preset value, the initial Lyapunov function obtained last time is used to solve the attraction domain of the power system to be analyzed to obtain an updated attraction domain estimation result. Based on the updated attraction domain estimation result, the optimal solution of the parameters of the grid controller is continued to be searched using pattern search and axial search until the step size is less than the second preset value to obtain the optimal parameter result.
[0177] In one embodiment, a dynamic analysis is performed on the dynamic model of the network controller to obtain a state vector equation, including:
[0178] The dynamic model of the grid controller is dynamically analyzed, and the state vector equations of the grid control power supply are obtained in the dynamic analysis. The state vector equations are:
[0179]
[0180] Among them, Δδ c is the controller equivalent relative phase angle, Δδ c =θ pll -θ ref , is the equivalent speed deviation of the controller, is the speed deviation of the system reference synchronous generator, is the q-axis voltage at the common coupling point where the grid-following control power source is connected to the grid, f n is the system rated frequency, is the rated speed of the system, K p and K i It is an adjustable parameter of the network controller.
[0181] The state vector equation of the reference synchronous generator is:
[0182]
[0183] in, is the relative speed of the reference synchronous generator, H ref is the reference synchronous generator moment of inertia, is the reference synchronous generator damping coefficient, is the reference synchronous generator mechanical power input, is the reference synchronous generator electrical power output.
[0184] The state vector equations of the remaining synchronous generators in the system are:
[0185]
[0186] Among them, δ i is the phase angle of the synchronous generator, is the relative speed of the synchronous generator, H i is the moment of inertia of the synchronous generator, is the damping coefficient of the synchronous generator, is the synchronous generator mechanical power input, The electrical power output of the synchronous generator.
[0187] In one embodiment, an extended inner domain method based on Lyapunov's second method is used in combination with the sum of squares theory to solve the attraction domain of the power system to be analyzed, and an estimation result of the attraction domain in the power system to be analyzed is obtained, including:
[0188] Set the criterion for the end of the loop γ 0 :=0,i:=1,j:=1, select the initial Lyapunov function, and use the attraction domain estimation problem model to solve it, where the attraction domain estimation problem model is:
[0189] find(V 0 ,λ1,λ3,s2)
[0190] st
[0191]
[0192] in, is the set of square sum polynomials, is a set of n-1 dimensional polynomial vectors, G is the equality constraint generated in the process of detrigonometric function of the system model, β is a positive number, r(z) is a positive definite polynomial, λ1 and λ3 are n-1 dimensional polynomial variables added to the equality constraint G in the positive point theorem, and s2 is the square sum polynomial variable added to the inequality constraint β-r(z) in the positive point theorem; and is a square sum polynomial,
[0193] set up Setting V (i) (z):=V (i-1)(z), γ=γ (i-1) , transform the attraction domain estimation problem into the first square sum planning problem, obtain the first square sum optimization problem model and solve it to obtain the first solution result, where the first square sum optimization problem is:
[0194] (SOSP1)max c
[0195] st
[0196]
[0197] in, is the set of square sum polynomials, is a set of n-1 dimensional polynomial vectors, G is the equality constraint generated in the process of detrigonometric function of the system model, β and c are positive numbers, p is a positive definite polynomial, λ2 and λ3 are n-1 dimensional polynomial variables added to the equality constraint G in the positive point theorem, s1, s2, s3 are square sum polynomial variables added to the inequality constraint in the positive point theorem, is a square sum polynomial, which is used to ensure that each constraint is semi-positive definite.
[0198] Taking the first solution result as a variable, the first square sum planning problem is transformed into the second square sum planning problem, and the second square sum optimization problem model is obtained and solved to obtain the second solution result, wherein the second square sum optimization problem model is:
[0199] (SOSP2)maxβ
[0200] st
[0201]
[0202] in, is the set of square sum polynomials, is a set of n-1 dimensional polynomial vectors, G is the equality constraint generated in the process of detrigonometric function of the system model, β and c are positive numbers, p is a positive definite polynomial, λ1, λ2, λ3 are n-1 dimensional polynomial variables added to the equality constraint G in the positive point theorem, s1, s2, s3 are square sum polynomial variables added to the inequality constraint in the positive point theorem, and is a square sum polynomial, which is used to ensure that each constraint is semi-positive definite.
[0203] The second solution results β and V(z) are stored as β (i) and V (i) (z), if β (i) -β(i-1) <∈ β , then if j>1 and and If the maximum absolute value of the coefficient is less than the preset criterion, the estimation result of the attraction domain and the current Lyapunov function are obtained. If not, V is set. (0) (z):=V (i) (z), P 0 (z):=V (i) (z), β (0) :=c (i) , i: = 1, j = j + 1, re-solve the first square sum optimization problem model and the second square sum optimization problem model until the attraction domain estimation result and the Lyapunov function are obtained;
[0204] If β (i) -β (i-1) >∈ β , otherwise set i=i+1, and re-solve the first sum of squares optimization problem model and the second sum of squares optimization problem model until the attraction domain estimation result and Lyapunov function are obtained.
[0205] In one embodiment, the calculation formula of the error index is:
[0206]
[0207] Among them, t ref represents the real critical removal time, t cal Represents the estimated critical resection time.
[0208] In one embodiment, the axial search is used to search for the optimal solution of the parameters of the network controller, including:
[0209] Along K p Direction search, if Then like Then Otherwise,
[0210] Along K i Direction search. Then like Then Otherwise, Among them, K p and K i is an adjustable parameter of the network controller, and d is the step size of the axial search.
[0211] In one embodiment, the optimal solution of the parameters of the network controller is searched by pattern search, including:
[0212] set up Along the y k =x k -x k-1 Direction search, if Δ(x k -αy k )<Δ(y k ), let x k+1 =x k -αy k Otherwise, let x k+1 =x k .
[0213] In one embodiment of the present application, a computer device is provided, which includes a memory and a processor, wherein a computer program is stored in the memory, and the above steps are implemented when the processor executes the computer program; the computer device provided in this embodiment has an implementation principle and technical effects similar to those of the above method embodiments, and will not be repeated here.
[0214] In one embodiment of the present application, a computer-readable storage medium is provided, on which a computer program is stored, and the above steps are implemented when the computer program is executed by a processor; the computer-readable storage medium provided in this embodiment has an implementation principle and technical effects similar to those of the above method embodiment, and will not be repeated here.
[0215] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
[0216] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0217] The specific embodiments described above further illustrate the purpose, technical solutions and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of protection of the present invention.
Claims
1. A method for optimizing network controller parameters, characterized in that: include: Constructing a dynamic model of a grid controller in the power system to be analyzed, and dynamically analyzing the dynamic model of the grid controller to obtain a state vector equation; The extended inner domain method based on Lyapunov's second method is used in combination with the sum of squares theory to solve the attraction domain of the power system to be analyzed, and the estimation result of the attraction domain in the power system to be analyzed is obtained; Performing a time domain simulation of the system fault state of the power system to be analyzed to obtain the real critical removal time and the state variable values at each moment, substituting the state variable values at each moment into the Lyapunov function to obtain multiple Lyapunov function values, selecting the moment corresponding to the Lyapunov function value that meets the preset conditions as the estimated critical removal time, and using the real critical removal time and the estimated critical removal time to obtain the error index of the attraction domain estimation result; If the error index is greater than the preset index, the optimal parameter search is performed based on the state vector equation using axial search. When the step size is less than the first preset value, the initial Lyapunov function obtained last time is used to solve the attraction domain of the power system to be analyzed, and an updated attraction domain estimation result is obtained. Based on the updated attraction domain estimation result, the optimal solution for the parameters of the grid controller is continued to be searched using pattern search and axial search until the step size is less than the second preset value, and the optimal parameter result is obtained.
2. The method for optimizing network controller parameters according to claim 1, characterized in that: The dynamic analysis of the dynamic model of the network controller is performed to obtain a state vector equation, including: The dynamic model of the grid controller is dynamically analyzed, and the state vector equations of the grid control power supply are obtained in the dynamic analysis. The state vector equations are: Among them, Δδ c is the controller equivalent relative phase angle, Δδ c =θ pll -θ ref , is the equivalent speed deviation of the controller, is the speed deviation of the system reference synchronous generator, is the q-axis voltage at the common coupling point where the grid-following control power source is connected to the grid, f n is the system rated frequency, is the rated speed of the system, K p and K i It is an adjustable parameter for the network controller; The state vector equation of the reference synchronous generator is: in, is the relative speed of the reference synchronous generator, H ref is the reference synchronous generator moment of inertia, is the reference synchronous generator damping coefficient, is the reference synchronous generator mechanical power input, is the reference synchronous generator electrical power output; The state vector equations of the remaining synchronous generators in the system are: Among them, δ i is the phase angle of the synchronous generator, is the relative speed of the synchronous generator, H i is the moment of inertia of the synchronous generator, is the damping coefficient of the synchronous generator, is the synchronous generator mechanical power input, The electrical power output of the synchronous generator.
3. The method for optimizing network controller parameters according to claim 1, characterized in that: The extended inner domain method based on Lyapunov's second method is used in combination with the sum of squares theory to solve the attraction domain of the power system to be analyzed, and the attraction domain estimation result in the power system to be analyzed is obtained, including: Set the criterion for the end of the loop γ 0 :=0,i:=1,j:=1, select the initial Lyapunov function, and use the attraction domain estimation problem model to solve it, where the attraction domain estimation problem model is: in, is the set of square sum polynomials, is a set of n-1 dimensional polynomial vectors, G is the equality constraint generated in the process of detrigonometric function of the system model, β is a positive number, r(z) is a positive definite polynomial, λ1 and λ3 are n-1 dimensional polynomial variables added to the equality constraint G in the positive point theorem, and s2 is the sum of squares polynomial variable added to the inequality constraint β-r(z) in the positive point theorem; and is a square sum polynomial, set up Setting V (i) (z):=V (i-1) (z),γ:=γ (i-1) , transform the attraction domain estimation problem into a first square sum planning problem, obtain a first square sum optimization problem model and solve it to obtain a first solution result, wherein the first square sum optimization problem is: in, is the set of square sum polynomials, is a set of n-1 dimensional polynomial vectors, G is the equality constraint generated in the process of detrigonometric function of the system model, β and c are positive numbers, p is a positive definite polynomial, λ2 and λ3 are n-1 dimensional polynomial variables added to the equality constraint G in the positive point theorem, s1, s2, s3 are square sum polynomial variables added to the inequality constraint in the positive point theorem, is a square sum polynomial, which is used to ensure that each constraint is semi-positive definite. Taking the first solution result as a variable, the first square sum planning problem is transformed into a second square sum planning problem, and a second square sum optimization problem model is obtained and solved to obtain a second solution result, wherein the second square sum optimization problem model is: in, is the set of square sum polynomials, is a set of n-1 dimensional polynomial vectors, G is the equality constraint generated in the process of detrigonometric function of the system model, β and c are positive numbers, p is a positive definite polynomial, λ1, λ2, λ3 are n-1 dimensional polynomial variables added to the equality constraint G in the positive point theorem, s1, s2, s3 are square sum polynomial variables added to the inequality constraint in the positive point theorem, and is a square sum polynomial, which is used to ensure that each constraint is semi-positive definite. The second solution results β and V(z) are stored as β (i) and V (i) (z), if β (i) -β (i-1) <∈ β , then if j>1 and and If the maximum absolute value of the coefficient is less than the preset criterion, the estimation result of the attraction domain and the current Lyapunov function are obtained. If not, V is set. (0) (z):=V (i) (z), P 0 (z):=V (i) (z), β (0) :=c (i) , i:=1, j=j+1, resolving the first sum-of-squares optimization problem model and the second sum-of-squares optimization problem model until the attraction domain estimation result and the Lyapunov function are obtained; If β (i) -β (i-1) >∈ β , otherwise set i=i+1, and re-solve the first sum of squares optimization problem model and the second sum of squares optimization problem model until the attraction domain estimation result and the Lyapunov function are obtained.
4. The method for optimizing network controller parameters according to claim 1, characterized in that: The calculation formula of the error index is: Among them, t ref represents the real critical removal time, t cal Represents the estimated critical resection time.
5. The method for optimizing network controller parameters according to claim 1, characterized in that: The method of searching for optimal parameters by using axial search includes: Along K p Direction search, if Then like Then Otherwise, Along K i Direction search. Then like Then Otherwise, Among them, K p and K i is an adjustable parameter of the network controller, and d is the step size of the axial search.
6. The method for optimizing network controller parameters according to claim 1, characterized in that: The method of searching for an optimal solution for the parameters of the network controller using a pattern search includes: set up Along the y k =x k -x k-1 Direction search, if Δ(x k -αy k )<Δ(y k ), let x k+1 =x k -αy k Otherwise, let x k+1 =x k .
7. A network controller parameter optimization device, characterized in that: include: A construction module is used to construct a dynamic model of a grid controller in the power system to be analyzed, and to dynamically analyze the dynamic model of the grid controller to obtain a state vector equation; An estimation module, used for solving the attraction domain of the power system to be analyzed by using an extended inner domain method based on Lyapunov's second method in combination with the sum of squares theory, and obtaining an estimation result of the attraction domain in the power system to be analyzed; An error determination module is used to perform a time domain simulation of the system fault state of the power system to be analyzed, obtain the real critical removal time and the state variable values at each moment, substitute the state variable values at each moment into the Lyapunov function, obtain multiple Lyapunov function values, select the moment corresponding to the Lyapunov function value that meets the preset conditions as the estimated critical removal time, and use the real critical removal time and the estimated critical removal time to obtain the error index of the attraction domain estimation result; A parameter calculation module is used to search for the optimal parameters based on the state vector equation using axial search if the error index is greater than a preset index. When the step size is less than a first preset value, the initial Lyapunov function obtained last time is used to solve the attraction domain of the power system to be analyzed to obtain an updated attraction domain estimation result. Based on the updated attraction domain estimation result, the optimal solution of the parameters of the grid controller is continued to be searched using pattern search and axial search until the step size is less than a second preset value to obtain the optimal parameter result.
8. The network controller parameter optimization device according to claim 7, characterized in that: The dynamic analysis of the dynamic model of the network controller is performed to obtain a state vector equation, including: The dynamic model of the grid controller is dynamically analyzed, and the state vector equations of the grid control power supply are obtained in the dynamic analysis. The state vector equations are: Among them, Δδ c is the controller equivalent relative phase angle, Δδ c =θ pll -θ ref , is the equivalent speed deviation of the controller, is the speed deviation of the system reference synchronous generator, is the q-axis voltage at the common coupling point where the grid-following control power source is connected to the grid, f n is the system rated frequency, is the rated speed of the system, K p and K i It is an adjustable parameter of the network controller. The state vector equation of the reference synchronous generator is: in, is the relative speed of the reference synchronous generator, H ref is the reference synchronous generator moment of inertia, is the reference synchronous generator damping coefficient, is the reference synchronous generator mechanical power input, is the reference synchronous generator electrical power output; The state vector equations of the remaining synchronous generators in the system are: Among them, δ i is the phase angle of the synchronous generator, is the relative speed of the synchronous generator, H i is the moment of inertia of the synchronous generator, is the damping coefficient of the synchronous generator, is the synchronous generator mechanical power input, The electrical power output of the synchronous generator.
9. A computer device, characterized in that: include: Memory for storing computer programs; A processor, configured to implement the network controller parameter optimization method as described in any one of claims 1 to 6 when executing the computer program.
10. A storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by the processor, the steps of the network controller parameter optimization method according to any one of claims 1 to 6 are implemented.
Citation Information
Patent Citations
Direct power regulation and control method for following network type distributed energy grid-connected system
CN114865720A
Power system attraction domain estimation method based on gravitational search algorithm
CN117010143A
