Method and device for evaluating transient stability of power system based on improved RNN (Recurrent Neural Network)

Abstract

The invention discloses an improved RNN-based power system transient stability evaluation method and device, and relates to the technical field of power system evaluation. According to the method, the transient stability boundary constraint can fully adapt to the characteristics of a power system through the rated operation data and the maximum angle-of-attack difference data of the generator, and through the rated operation data and the maximum angle-of-attack difference data which are convenient to collect, the construction process of the transient stability boundary constraint is simplified, and the calculation complexity is reduced; and finally, the solving efficiency and accuracy of the transient stability boundary are improved. By combining the accurate fluctuation data of the power system with the transient stability evaluation model, the transient stability-related characteristics in the fluctuation data can be fully and quickly mined, the accuracy of transient stability evaluation is improved, and the low efficiency and the low cost caused by a large amount of direct calculation are avoided. Therefore, the evaluation efficiency and accuracy of transient stability evaluation of the power system are finally improved, and safe and stable operation of the power system is ensured.

Description

Technical Field

[0001] The present invention belongs to the technical field of power system evaluation, and in particular relates to a method and device for evaluating transient stability of a power system based on an improved RNN. Background Art

[0002] With the continuous growth of electricity demand, the current power system has been running closer and closer to its physical limit. In the case of difficulty in expanding the system at the power generation and transmission ends, how to dynamically evaluate the stability of the power system has become a key issue to ensure the safe operation of the power system. In the stability assessment of the power system, transient stability is the most critical one in the power system. It is defined as the ability of the power system to maintain synchronization when it is subject to large disturbances such as load loss, generator loss or transmission system facility failure. Among them, how to construct a transient stability safety boundary is a necessary prerequisite for realizing online transient stability assessment of the power system.

[0003] At present, in the transient stability assessment technology of power systems, time domain simulation and direct method are usually used for transient stability assessment, and transient stability boundaries are determined by numerical integration method, energy function method and trajectory analysis method. However, the time domain simulation has a large amount of calculation and is time-consuming, and can only be roughly judged based on engineering experience methods such as the oscillation trend of the power angle curve. The direct method has a complex solution process, which makes the current transient stability assessment of power systems inefficient and inaccurate. The common problems of the numerical integration method, the energy function method and the trajectory analysis method are poor adaptability and complex solution process, and they cannot adapt to the increasing scale of the current power system, which makes the solution efficiency and accuracy of the transient stability boundary low, resulting in low efficiency and low accuracy of transient stability assessment of power systems. Therefore, there is an urgent need for a new method and device for transient stability assessment of power systems to solve the defects of the existing technology. Summary of the invention

[0004] The present invention aims to provide a method and device for transient stability assessment of power system based on improved RNN to solve the above technical problems. The transient stability boundary constraints are constructed and fluctuation data is collected through the rated operation data of the power system and the maximum angle of attack difference data of the generator. The transient stability assessment model is combined to realize the transient stability assessment of the power system, thereby improving the assessment efficiency and accuracy of the transient stability assessment of the power system.

[0005] In order to solve the above technical problems, an embodiment of the present invention provides a method for evaluating transient stability of a power system based on an improved RNN, comprising:

[0006] Constructing initial transient stability boundary constraints based on rated operation data of the power system;

[0007] Collecting maximum angle of attack difference data of generators in the power system, and constructing a first transient stability boundary constraint in combination with the initial transient stability boundary constraint;

[0008] When the power system is disturbed, collecting fluctuation data of the power system according to the first transient stability boundary constraint to obtain a fluctuation data set;

[0009] A power system transient stability assessment model is constructed, and the fluctuation data set is input into the power system transient stability assessment model.

[0010] It can be understood that, compared with the prior art, the present invention constructs the initial transient stability boundary constraint through the rated operation data of the power system, and then builds the first transient stability boundary constraint in combination with the maximum angle of attack difference data of the generator in the power system, so that the transient stability boundary constraint can fully adapt to the characteristics of the power system by using the rated operation data and the maximum angle of attack difference data of the generator that are closely related to the safe and stable operation of the power system, and simplifies the construction process of the transient stability boundary constraint and reduces the calculation complexity through the convenient collection of rated operation data and maximum angle of attack difference data, and finally improves the efficiency and accuracy of solving the transient stability boundary. Through the accurate transient stability boundary, the fluctuation data of the power system under large disturbance can be collected, and then the transient stability evaluation result of the power system is obtained by combining the transient stability evaluation model. By combining the accurate power system fluctuation data with the transient stability evaluation model, compared with the existing time domain simulation and direct method, the characteristics of transient stability in the fluctuation data can be fully and quickly mined, the accuracy of transient stability evaluation is improved, and the inefficiency caused by a large number of direct calculations is avoided, thereby finally improving the evaluation efficiency and accuracy of transient stability evaluation of the power system, and ensuring the safe and stable operation of the power system.

[0011] As a preferred solution, the initial transient stability boundary constraint is constructed according to the rated operation data of the power system, specifically including:

[0012] The rated operation data of the power system include: active power of the generator, reactive power of the generator, voltage amplitude of the node and phase angle of the node;

[0013] Constructing a first variable equation group and a second variable equation group of the power system according to the active power of the generator, the reactive power of the generator, the voltage amplitude of the node and the phase angle of the node;

[0014] The power flow equation and the initial state equation of the power system are obtained, and the initial transient stability boundary constraints are constructed by combining the first variable equation group and the second variable equation group.

[0015] This preferred scheme constructs the first variable equation group and the second variable equation group of the power system, and constructs the initial transient stability boundary constraints in combination with the power system's power flow equation and the initial state equation. It can study the transient stability boundary of the power system, thereby obtaining more accurate power data at the transient stability boundary, providing accurate data support for the transient stability assessment of the power system, and improving the accuracy of the transient stability assessment results of the power system.

[0016] As a preferred solution, the collecting of the maximum angle of attack difference data of the generator in the power system and the construction of the first transient stability boundary constraint in combination with the initial transient stability boundary constraint specifically include:

[0017] According to a preset time window length, the maximum angle of attack difference data of the generator in the power system is collected to obtain the maximum angle of attack difference data;

[0018] According to the maximum angle of attack difference data, a second transient stability boundary constraint is constructed in combination with the initial transient stability boundary constraint; wherein the second transient stability boundary constraint specifically includes:

[0019]

[0020] Where i = 1, 2, 3, ..., N G , N G represents the number of generators in the power system, T represents the length of the simulation time window; δ max The maximum power angle limit to maintain system stability; It represents the absolute power angle value of the i-th generator at time t under the operating state u and the k-th expected fault; Indicates the power angle value in the central inertial coordinate system; T Ji represents the inertia time constant of the i-th generator;

[0021] The second transient stability boundary constraint is converted to obtain a first transient stability boundary constraint.

[0022] This preferred solution constructs the initial transient stability boundary constraint through the rated operation data of the power system, and then constructs the first transient stability boundary constraint in combination with the maximum angle of attack difference data of the generator in the power system. The rated operation data and the maximum angle of attack difference data of the generator, which are closely related to the safe and stable operation of the power system, enable the transient stability boundary constraint to fully adapt to the characteristics of the power system. The rated operation data and the maximum angle of attack difference data, which are easily collected, simplify the construction process of the transient stability boundary constraint and reduce the calculation complexity, thereby ultimately improving the efficiency and accuracy of solving the transient stability boundary.

[0023] As a preferred solution, the first transient stability boundary constraint specifically includes:

[0024]

[0025] Among them, Φ k (u) is used to represent the dynamic performance index of the power system at a given operating point and expected faults; θ k represents the weight function, which is used to measure whether the system is in a stable state during the fault period, δ k (t|u) represents the dynamic state variables of the system as a function of time when the kth expected fault occurs, and depends on the given static variable u. It represents the power angle value in the central inertial coordinate system; It represents the absolute power angle value of the i-th generator at time t under the operating state u and the k-th expected fault. Λ is a binary constant used to distinguish whether the system is stable or not. StableOP means the system is stable, and UnstableOP means the system is unstable.

[0026] This preferred solution constructs a first transient stability boundary constraint based on the initial transient stability boundary constraint and combined with the maximum angle of attack difference data of the generator in the power system, so that the obtained transient stability boundary constraint for the power system is more in line with the actual operation of the power system, thereby being able to obtain more accurate power data at the transient stability boundary, providing accurate data support for the transient stability assessment of the power system, and improving the accuracy of the transient stability assessment results of the power system.

[0027] As a preferred solution, when the power system is disturbed, the fluctuation data of the power system is collected according to the first transient stability boundary constraint to obtain a fluctuation data set, which specifically includes:

[0028] Obtaining a data acquisition direction according to the first transient stability boundary constraint;

[0029] Obtaining a distance between a current operating state of the power system and a transient stable operating state according to the first transient stability boundary constraint;

[0030] The step length coefficient is determined according to the distance between the current operation state of the power system and the transient stable operation state, and then the search step length is obtained;

[0031] When the power system is disturbed, the fluctuation data of the power system is collected according to the data collection direction, the distance between the current operating state of the power system and the transient stable operating state, and the search step size until the power system is in a transient stable operating state, thereby obtaining a fluctuation data set.

[0032] This preferred scheme obtains more accurate data samples close to the transient stability boundary through data acquisition direction, the distance between the current operating state of the power system and the transient stability operating state, and the search step size, providing accurate data support for the transient stability assessment of the power system and improving the accuracy of the transient stability assessment results of the power system.

[0033] As a preferred solution, obtaining the data acquisition direction according to the first transient stability boundary constraint specifically includes:

[0034] Solving the first-order derivative of the first transient stability boundary constraint to obtain a first-order derivative function of the stability index;

[0035] The data collection direction is obtained according to the first-order derivative function of the stability indicator.

[0036] This preferred solution can ensure the accuracy of the data samples collected close to the transient stability boundary by determining the data collection direction, provide accurate data support for the transient stability assessment of the power system, and improve the accuracy of the transient stability assessment results of the power system.

[0037] As a preferred solution, the step coefficient is determined according to the distance between the current operating state of the power system and the transient stable operating state, and then the search step is obtained, which specifically includes:

[0038] Obtaining the interval of the distance between the current operating state of the power system and the transient stable operating state according to the preset distance interval set;

[0039] Determine the step size coefficient according to the interval;

[0040] A vector value of the current operating state of the current power system is obtained, and the search step length is determined in combination with the step length coefficient.

[0041] This preferred solution can ensure the accuracy of the data samples collected close to the transient stability boundary by determining the search step size, provide accurate data support for the transient stability assessment of the power system, and improve the accuracy of the transient stability assessment results of the power system.

[0042] As a preferred solution, the constructing of the power system transient stability assessment model, inputting the fluctuation data set into the power system transient stability assessment model, and obtaining the transient stability assessment result of the power system specifically includes:

[0043] Construct a power system transient stability assessment model according to a preset model construction method;

[0044] Inputting the fluctuation data set into the power system transient stability assessment model to obtain a transient stability assessment value;

[0045] According to the preset transient stability assessment threshold, the transient stability assessment result of the power system is obtained in combination with the transient stability assessment value.

[0046] By combining accurate power system fluctuation data with a transient stability assessment model, this preferred solution can fully and quickly mine the characteristics of transient stability in fluctuation data compared to existing time-domain simulation and direct methods, thereby improving the accuracy of transient stability assessment and avoiding the inefficiency caused by a large number of direct calculations. This ultimately improves the assessment efficiency and accuracy of power system transient stability assessment and ensures safe and stable operation of the power system.

[0047] As a preferred solution, the power system transient stability assessment model specifically includes:

[0048] The power system transient stability assessment model comprises: a first layer neural network, a second layer neural network and an output layer;

[0049] The output end of the first neural network is connected to the input end of the second neural network, and the output end of the second neural network is connected to the input end of the output layer; the input end of the first neural network is used to receive the fluctuation data set of the power system transient stability assessment model; the output layer is used to output the transient stability assessment value of the power system transient stability assessment model;

[0050] Wherein, the first layer of neural network is a dynamically stable recursive neural network, which is used to classify the fluctuation data set input into the power system transient stability assessment model;

[0051] The hidden state expression of the dynamically stable recurrent neural network is:

[0052]

[0053] in, for representing the hidden state of the dynamically stable recurrent neural network; for representing an input of the dynamically stable recurrent neural network; is the jump coefficient; is a diagonal matrix with α as the diagonal entry i (j) , is the activation function; the parameter m∈N is used to indicate the number of weighted skip connections across time steps;

[0054] The second layer of neural network is a fully connected layer, which is used to reduce the dimension of the output data of the first layer of neural network until the required dimension of the output layer is reached;

[0055] The output layer is used to classify and analyze the output data of the fully connected layer to obtain a transient stability assessment value.

[0056] By combining accurate power system fluctuation data with a transient stability assessment model, this preferred solution can fully and quickly mine the characteristics of transient stability in fluctuation data compared to existing time-domain simulation and direct methods, thereby improving the accuracy of transient stability assessment and avoiding the inefficiency caused by a large number of direct calculations. This ultimately improves the assessment efficiency and accuracy of power system transient stability assessment and ensures safe and stable operation of the power system.

[0057] Accordingly, an embodiment of the present invention provides a power system transient stability assessment device based on an improved RNN, comprising: an initial transient stability boundary constraint construction module, a first transient stability boundary constraint construction module, a fluctuation data acquisition module, and a transient stability assessment module;

[0058] Wherein, the initial transient stability boundary constraint construction module is used to construct the initial transient stability boundary constraint according to the rated operation data of the power system;

[0059] The first transient stability boundary constraint building module is used to collect the maximum angle of attack difference data of the generator in the power system, and to build the first transient stability boundary constraint in combination with the initial transient stability boundary constraint;

[0060] The fluctuation data collection module is used to collect the fluctuation data of the power system according to the first transient stability boundary constraint when the power system is disturbed, so as to obtain a fluctuation data set;

[0061] The transient stability assessment module is used to construct a transient stability assessment model for the power system, input the fluctuation data set into the transient stability assessment model for the power system, and obtain a transient stability assessment result for the power system.

[0062] It can be understood that, compared with the prior art, the present device constructs the initial transient stability boundary constraint through the rated operation data of the power system, and then constructs the first transient stability boundary constraint in combination with the maximum angle of attack difference data of the generator in the power system, so that the transient stability boundary constraint can fully adapt to the characteristics of the power system by using the rated operation data and the maximum angle of attack difference data of the generator that are closely related to the safe and stable operation of the power system, and simplifies the construction process of the transient stability boundary constraint and reduces the calculation complexity through the convenient collection of rated operation data and maximum angle of attack difference data, and finally improves the efficiency and accuracy of solving the transient stability boundary. Through the accurate transient stability boundary, the fluctuation data of the power system under large disturbance can be collected, and then the transient stability evaluation result of the power system is obtained by combining the transient stability evaluation model. By combining the accurate power system fluctuation data with the transient stability evaluation model, compared with the existing time domain simulation and direct method, it can fully and quickly mine the characteristics of transient stability in the fluctuation data, improve the accuracy of transient stability evaluation, avoid the inefficiency caused by a large number of direct calculations, and finally improve the evaluation efficiency and accuracy of transient stability evaluation of the power system, and ensure the safe and stable operation of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 :A flowchart of a method for evaluating transient stability of a power system based on an improved RNN provided in an embodiment of the present invention;

[0064] Figure 2 : A structural diagram of a transient stability assessment model based on a DSRNN model provided in an embodiment of the present invention;

[0065] Figure 3 :A flowchart of another method for evaluating transient stability of a power system based on an improved RNN provided in an embodiment of the present invention;

[0066] Figure 4 : A schematic diagram of the structure of a power system transient stability assessment device based on an improved RNN provided in an embodiment of the present invention;

[0067] Among them, 201: initial transient stability boundary constraint construction module; 202: first transient stability boundary constraint construction module; 203: fluctuation data acquisition module; 204: transient stability assessment module. DETAILED DESCRIPTION

[0068] 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.

[0069] With the continuous growth of power industry demand, power systems are increasingly forced to operate close to their physical limits, but expanding power generation and transmission systems is usually difficult and expensive. Therefore, dynamic security assessment has become a key issue to ensure the safe operation of the system. Among various stability definitions, transient stability has always been the traditional dominant stability issue for most power systems. Transient stability is defined as the ability of a power system to maintain synchronization when subjected to large disturbances, such as load loss, generator loss, or transmission system facility failure. The boundary problem of power system transient stability is the core content of power system stability assessment and one of the core issues of the new generation of power grid intelligent control system. Constructing a transient stability safety boundary is a necessary prerequisite for realizing online transient stability assessment of power systems. Using the constructed stability boundary, not only can stability conclusions be drawn based on the current real-time operation and disturbance information of the power grid, but also preventive or emergency control strategies can be given based on this. Existing transient stability assessment methods often only focus on transient stability or transient instability under normal conditions, while ignoring the critical stability state of the power grid, or cannot distinguish the critical stability state of the power grid, resulting in a fuzzy area in the power grid security assessment. Therefore, the boundary problem is one of the biggest problems restricting the improvement of transient stability identification accuracy. At present, relatively time-consuming time domain simulation and direct method are still usually used for transient stability assessment in engineering practice. The main methods that can currently solve transient stability boundary problems include numerical integration method, energy function method and trajectory analysis method.

[0070] With the increasing integration of renewable energy, the safe and stable operation of the power system has ushered in new challenges, especially the operational challenges brought about by the reduction of system inertia, which will have a significant impact on the assessment of power system transient stability. With the development of new generation technologies, the traditional problem of transient stability assessment has become more complex, and its importance has become increasingly prominent. It has become the focus of power system technology research. Traditional power transient stability assessment methods, such as direct methods and time domain simulation methods, can no longer meet today's rapidly changing technological needs, and their defects are becoming increasingly prominent.

[0071] The time domain simulation method has the following three limitations to a certain extent: first, the method has a large amount of calculation, is time-consuming, and has low calculation efficiency; second, although the algorithm can obtain the time solution of the differential algebraic equations of the disturbed system through time domain simulation, there is no unified and clear stability criterion available for use, and it can only be roughly judged based on engineering experience methods such as the oscillation trend of the power angle curve or the maximum relative power angle difference within a certain time window; third, this method cannot obtain mechanism indicators such as stability margin. The direct method mainly starts from the energy function and then obtains the critical energy value or critical fixed value, but its model adaptability is poor and the calculation process is complicated. The common problems of the numerical integration method, the energy function method and the trajectory analysis method are that the model adaptability is poor and the calculation process is complicated. At the same time, the former has a large error due to its conservativeness, and the latter will become more complicated as the scale of the power grid increases.

[0072] As described above, in order to solve the defects of the above-mentioned prior art, an embodiment of the present invention provides a method and device for transient stability assessment of a power system based on an improved RNN, and the specific content is described below.

[0073] Embodiment 1

[0074] Please refer to Figure 1 , which is a flowchart of a method for transient stability assessment of a power system based on an improved RNN provided in an embodiment of the present invention, including steps S101 to S104.

[0075] Step S101: constructing initial transient stability boundary constraints according to rated operation data of the power system.

[0076] In this embodiment, the initial transient stability boundary constraint is constructed according to the rated operation data of the power system, specifically including:

[0077] The rated operation data of the power system include: active power of the generator, reactive power of the generator, voltage amplitude of the node and phase angle of the node;

[0078] Constructing a first variable equation group and a second variable equation group of the power system according to the active power of the generator, the reactive power of the generator, the voltage amplitude of the node and the phase angle of the node;

[0079] The power flow equation and the initial state equation of the power system are obtained, and the initial transient stability boundary constraints are constructed by combining the first variable equation group and the second variable equation group.

[0080] This embodiment constructs the first variable equation group and the second variable equation group of the power system, and constructs the initial transient stability boundary constraints in combination with the power system's power flow equation and the initial state equation. It can study the transient stability boundary of the power system, thereby obtaining more accurate power data at the transient stability boundary, providing accurate data support for the transient stability assessment of the power system, and improving the accuracy of the transient stability assessment results of the power system.

[0081] In an optional embodiment, the first variable equation group of the power system refers to the differential equation group F k The second variable equation group of the power system refers to the algebraic equation group G k .

[0082] The first variable equation system and the second variable equation system are as follows:

[0083]

[0084] In formula (1), u is the static variable of the power system at a given operating point (i.e., a given operating state, which is also the same in the following description), including the active and reactive power of the generator, the voltage amplitude and phase angle of each node; x0 is the initial value of the state variable, x k (t) is the time-varying state, y k (t) is the time-varying operating variable; x k and k are time-varying state and operational variables, respectively.

[0085] In an optional embodiment, the initial transient stability boundary constraint includes:

[0086]

[0087] In formula (2), P represents the power flow equation, J represents the system steady-state operation constraint, Z is the system initial state equation, Φ k is the system transient stability constraint under the kth expected fault; x0 is the initial value of the state variable, u is the static variable of the power system at a given operating point, including the active and reactive power of the generator, the voltage amplitude and phase angle of each node; J and They represent the lower and upper bounds of the system’s steady-state operation constraints respectively; Φ and They represent the lower and upper bounds of the system transient stability constraints respectively.

[0088] It should be noted that the time domain solution can be obtained by combining equations (1) and (2). However, in the time domain solution process, only the system transient stability constraint is time-consuming, which affects the feasibility of constructing the transient stability boundary. Therefore, the embodiment of the present invention only focuses on the transient part of the constraint, and constructs the transient stability constraint by adopting the maximum power angle difference of each generator within a given time window, as described in step S102.

[0089] Step S102: collecting maximum angle of attack difference data of the generator in the power system, and constructing a first transient stability boundary constraint in combination with the initial transient stability boundary constraint.

[0090] In this embodiment, the maximum angle of attack difference data of the generator in the power system is collected, and the first transient stability boundary constraint is constructed in combination with the initial transient stability boundary constraint, specifically including:

[0091] According to a preset time window length, the maximum angle of attack difference data of the generator in the power system is collected to obtain the maximum angle of attack difference data;

[0092] According to the maximum angle of attack difference data, a second transient stability boundary constraint is constructed in combination with the initial transient stability boundary constraint; wherein the second transient stability boundary constraint specifically includes:

[0093]

[0094] In formula (3) and formula (4), i = 1, 2, 3, ..., N G , N G represents the number of generators in the power system, T represents the length of the simulation time window; δ max The maximum power angle limit to maintain system stability; It represents the absolute power angle value of the i-th generator at time t under the operating state u and the k-th expected fault; Indicates the power angle value in the central inertial coordinate system; T Ji represents the inertia time constant of the i-th generator;

[0095] The second transient stability boundary constraint is converted to obtain a first transient stability boundary constraint.

[0096] It should be noted that δ max In order to maintain the maximum power angle limit for system stability, this value is a preset value. The typical value is generally between 100° and 180°. The higher the value, the higher the flexibility of system operation, but it is more likely to cause missed reporting errors (i.e., judging unstable faults as stable faults). The present invention sets the value to 110°.

[0097] This embodiment constructs an initial transient stability boundary constraint through the rated operation data of the power system, and then constructs the first transient stability boundary constraint in combination with the maximum angle of attack difference data of the generator in the power system. The rated operation data and the maximum angle of attack difference data of the generator, which are closely related to the safe and stable operation of the power system, enable the transient stability boundary constraint to fully adapt to the characteristics of the power system. The rated operation data and the maximum angle of attack difference data, which are conveniently collected, simplify the construction process of the transient stability boundary constraint and reduce the calculation complexity, thereby ultimately improving the efficiency and accuracy of solving the transient stability boundary.

[0098] In this embodiment, the first transient stability boundary constraint specifically includes:

[0099]

[0100] Among them, v k (u) is used to represent the dynamic performance index of the power system at a given operating point and expected faults; θ k represents the weight function, which is used to measure whether the system is in a stable state during the fault period, δ k (t|u) represents the dynamic state variables of the system as a function of time when the kth expected fault occurs, and depends on the given static variables u (such as generator power, node voltage, etc.). It represents the power angle value in the central inertial coordinate system; It represents the absolute power angle value of the i-th generator at time t under the operating state u and the k-th expected fault. Λ is a binary constant used to distinguish whether the system is stable or not. StableOP means the system is stable, and UnstableOP means the system is unstable.

[0101] It should be noted that, by combining equations (5), (6) and (7), a power system dynamic performance index Φ for a given operating point and expected fault combination is defined: k (u). In essence, this indicator can be considered as the maximum power angle curve, the straight line δ = δ max , the area enclosed by the straight line t=0 and the straight line t=T to measure the distance between the sample point and the transient stability boundary.

[0102] This embodiment constructs a first transient stability boundary constraint based on the initial transient stability boundary constraint and combined with the maximum angle of attack difference data of the generator in the power system, so that the obtained transient stability boundary constraint for the power system is more in line with the actual operation of the power system, thereby being able to obtain more accurate power data at the transient stability boundary, providing accurate data support for the transient stability assessment of the power system, and improving the accuracy of the transient stability assessment results of the power system.

[0103] It should be noted that in order to obtain data samples from the power system being disturbed to the transient stability boundary, three pieces of information need to be determined, namely, the data collection direction, the distance between the current operating state of the power system and the transient stability operating state, and the search step. The specific determination method is described in the following step S103.

[0104] Step S103: when the power system is disturbed, the fluctuation data of the power system is collected according to the first transient stability boundary constraint to obtain a fluctuation data set.

[0105] In this embodiment, when the power system is disturbed, the fluctuation data of the power system is collected according to the first transient stability boundary constraint to obtain the fluctuation data set, which specifically includes:

[0106] Obtaining a data acquisition direction according to the first transient stability boundary constraint;

[0107] Obtaining a distance between a current operating state of the power system and a transient stable operating state according to the first transient stability boundary constraint;

[0108] The step length coefficient is determined according to the distance between the current operation state of the power system and the transient stable operation state, and then the search step length is obtained;

[0109] When the power system is disturbed, the fluctuation data of the power system is collected according to the data collection direction, the distance between the current operating state of the power system and the transient stable operating state, and the search step size until the power system is in a transient stable operating state, thereby obtaining a fluctuation data set.

[0110] This embodiment obtains more accurate data samples close to the transient stability boundary through data collection direction, the distance between the current operating state of the power system and the transient stability operating state, and the search step size, thereby providing accurate data support for the transient stability assessment of the power system and improving the accuracy of the transient stability assessment results of the power system.

[0111] It should be noted that the distance between the current operating state and the transient stable operating state of the power system in the embodiment of the present invention is expressed by Φ k (u) to indicate.

[0112] In this embodiment, obtaining the data collection direction according to the first transient stability boundary constraint specifically includes:

[0113] Solving the first-order derivative of the first transient stability boundary constraint to obtain a first-order derivative function of the stability index;

[0114] The data collection direction is obtained according to the first-order derivative function of the stability indicator.

[0115] This embodiment can ensure the accuracy of the data samples collected close to the transient stability boundary by determining the data collection direction, provide accurate data support for the transient stability assessment of the power system, and improve the accuracy of the transient stability assessment results of the power system.

[0116] In an optional embodiment, performing a first-order derivative solution on the first transient stability boundary constraint to obtain a first-order derivative function of a stability index specifically includes:

[0117]

[0118] H k [u,x k (t),y k (t),θ k (t),β k (t)]=

[0119] θ k +(λ k ) T F k (u,x k ,y k )+(β k ) T G k (u,x k ,y k ) (9),

[0120]

[0121] In formula (8), is the first-order derivative function of the stability index, Φ k (u) is a dynamic performance index used to represent the power system at a given operating point and expected faults. u is a static variable of the power system at a given operating point, including the active and reactive power of the generator, the voltage amplitude and phase angle of each node, and H k [u,x k (t),y k (t),θ k (t),β k (t)] is the Hamiltonian function corresponding to the transient stability constraint of the kth anticipated fault;

[0122] In formula (9) and formula (10), H k [u,x k (t),y k (t),θ k (t),β k (t)] is the Hamiltonian function corresponding to the transient stability constraint of the kth anticipated fault, θ krepresents the weight function, which is used to measure whether the system is in a stable state during the fault period, λ k and β k is the time solution of the state equation of formula (10), λ k (T) is the terminal condition of the adjoint equation, λ k (t) represents the Lagrange multiplier at time t, which describes the system state λ k (t) sensitivity to the objective function Φk(u); (λ k ) T represents the transpose of the Lagrange multiplier, combined with the differential equation F k , used to describe the impact of the system state on the objective function; β k (t) corresponds to the algebraic equation G k The Lagrange multiplier of (β k ) T Is β k The transposed matrix of (t); x k (t) is the time-varying state, y k (t) is the time-varying operating variable, x k and k are time-varying state and operating variables respectively; u is the static variable of the power system at a given operating point, including the active and reactive power of the generator, the voltage amplitude and phase angle of each node.

[0123] In this embodiment, the step size coefficient is determined according to the distance between the current operating state of the power system and the transient stable operating state, and then the search step size is obtained, which specifically includes:

[0124] Obtaining the interval of the distance between the current operating state of the power system and the transient stable operating state according to the preset distance interval set;

[0125] Determine the step size coefficient according to the interval;

[0126] A vector value of the current operating state of the current power system is obtained, and the search step length is determined in combination with the step length coefficient.

[0127] This embodiment can ensure the accuracy of the data samples collected close to the transient stability boundary by determining the search step length, provide accurate data support for the transient stability assessment of the power system, and improve the accuracy of the transient stability assessment results of the power system.

[0128] In an optional embodiment, the process of determining the step size coefficient specifically includes:

[0129]

[0130] In formula (11), v i,jis the step coefficient, which is the distance Φ between the current operating point and the transient stability boundary k (u) a preset coefficient associated with it, is a scalar, and the distance satisfy Since the power system is a highly nonlinear system, the exact step size required to search for the transient stability boundary cannot be predetermined as a constant. Therefore, as the operating point approaches the stability boundary, the coefficient v i,j The value of should also be reduced accordingly.

[0131] In an optional embodiment, the obtaining of the vector value of the current operating state of the current power system and determining the search step length in combination with the step length coefficient specifically includes:

[0132]

[0133] In formula (12), v i,j is the step size coefficient, is the search step length, u max It is a vector composed of the maximum values ​​of all controllable variables at the current operating point.

[0134] In an optional embodiment, collecting fluctuation data of the power system specifically includes:

[0135] Get the operating point (operating state) in the power that needs to be sampled and simulated in the time domain:

[0136]

[0137] In formula (13), u m+1 represents the value of the control variable (operating point) at the m+1th iteration, u m represents the value of the control variable (operating point) at the mth iteration, is the search step length, is the first-order derivative function of the stability index.

[0138] Step S104: constructing a power system transient stability assessment model, inputting the fluctuation data set into the power system transient stability assessment model, and obtaining a transient stability assessment result of the power system.

[0139] In this embodiment, the constructing of the power system transient stability assessment model, inputting the fluctuation data set into the power system transient stability assessment model, and obtaining the transient stability assessment result of the power system specifically includes:

[0140] Construct a power system transient stability assessment model according to a preset model construction method;

[0141] Inputting the fluctuation data set into the power system transient stability assessment model to obtain a transient stability assessment value;

[0142] According to the preset transient stability assessment threshold, the transient stability assessment result of the power system is obtained in combination with the transient stability assessment value.

[0143] By combining accurate power system fluctuation data with a transient stability assessment model, this embodiment can fully and quickly mine the characteristics of transient stability in fluctuation data compared to existing time-domain simulation and direct methods, thereby improving the accuracy of transient stability assessment and avoiding the inefficiency caused by a large number of direct calculations, thereby ultimately improving the assessment efficiency and accuracy of power system transient stability assessment and ensuring the safe and stable operation of the power system.

[0144] In this embodiment, the power system transient stability assessment model specifically includes:

[0145] The power system transient stability assessment model comprises: a first layer neural network, a second layer neural network and an output layer;

[0146] The output end of the first neural network is connected to the input end of the second neural network, and the output end of the second neural network is connected to the input end of the output layer; the input end of the first neural network is used to receive the fluctuation data set of the power system transient stability assessment model; the output layer is used to output the transient stability assessment value of the power system transient stability assessment model;

[0147] Wherein, the first layer of neural network is a dynamically stable recursive neural network, which is used to classify the fluctuation data set input into the power system transient stability assessment model;

[0148] The hidden state expression of the dynamically stable recurrent neural network is:

[0149]

[0150] in, for representing the hidden state of the dynamically stable recurrent neural network; for representing an input of the dynamically stable recurrent neural network; is the jump coefficient; is a diagonal matrix with α as the diagonal entry i (j) , is the activation function; the parameter m∈N is used to indicate the number of weighted skip connections across time steps;

[0151] The second layer of neural network is a fully connected layer, which is used to reduce the dimension of the output data of the first layer of neural network until the required dimension of the output layer is reached;

[0152] The output layer is used to classify and analyze the output data of the fully connected layer to obtain a transient stability assessment value.

[0153] In an optional embodiment, the dynamically stabilized recurrent neural network model (DSRNN) includes a parameter m∈N, which represents the number of weighted skip connections across time steps, as shown in the following formula (14):

[0154]

[0155] In formula (14), is the jump coefficient; is a diagonal matrix with diagonal entries α i (j) , is the activation function; is the hidden state of the network; For input.

[0156] This embodiment achieves periodic hidden state stability by dynamically stabilizing the recursive neural network and adjusting the weights of the time jump connections, thereby alleviating the problems of gradient vanishing and explosion, making it more computationally stable and efficient, thereby improving the efficiency and accuracy of transient stability assessment of the power system.

[0157] In an alternative embodiment, please refer to Figure 2 , is a structural diagram of a transient stability assessment model based on a DSRNN model provided by an embodiment of the present invention. The original features of the model are (x1, x2, …, x t ), when the fault disappears, the phasor measurement unit (PMU) inputs the data collected at the first moment into x1. In order to evaluate the accuracy of the model, a two-layer neural network is constructed at a certain moment: the first layer uses DSRNN as a classifier for transient stability assessment. The second layer uses a Dense layer, that is, a fully connected layer, the purpose of which is to reduce the dimension of the DSRNN output data until it reaches the dimension required by the output layer. Then, by using the Sigmoid function to classify and analyze the evaluation results, the PMU collects the data at the next moment and compares it with the memory information at the previous moment to obtain the transient stability evaluation result at that moment. If the evaluation time is consistent with the set value, the output The evaluation results at all times are in is the value of Sigmoid, and its range is (0,1). Sigmoid is the activation function of the output layer of the binary classification problem. It usually has a threshold δ=0.5, and its stability criterion is shown in the following formula (15):

[0158]

[0159] The transient stability assessment result of the power system is obtained according to the transient stability assessment value output by the transient stability assessment model.

[0160] In an alternative embodiment, please refer to Figure 3 , which is a flow chart of another method for transient stability assessment of a power system based on an improved RNN provided by an embodiment of the present invention; first, the transient stability definition is studied based on the mathematical mechanism of the transient stability boundary; then, the transient stability constraint is constructed by using the maximum common angle difference of each generator within a given time window, and a large amount of transient stability boundary data samples are obtained, and a transient stability assessment model based on an improved RNN (Recurrent Neural Network) is constructed, and the transient stability assessment result of the power system is obtained through the transient stability assessment model.

[0161] This embodiment constructs the initial transient stability boundary constraint through the rated operation data of the power system, and then builds the first transient stability boundary constraint in combination with the maximum angle of attack difference data of the generator in the power system. The rated operation data and the maximum angle of attack difference data of the generator that are closely related to the safe and stable operation of the power system are used to make the transient stability boundary constraint fully adapt to the characteristics of the power system, and the rated operation data and the maximum angle of attack difference data that are conveniently collected simplify the construction process of the transient stability boundary constraint and reduce the calculation complexity, and finally improve the efficiency and accuracy of the transient stability boundary solution. Through the accurate transient stability boundary, the fluctuation data of the power system under large disturbance can be collected, and then the transient stability evaluation result of the power system is obtained in combination with the transient stability evaluation model. By combining the accurate power system fluctuation data with the transient stability evaluation model, compared with the existing time domain simulation and direct method, the characteristics of transient stability in the fluctuation data can be fully and quickly mined, the accuracy of transient stability evaluation is improved, and the inefficiency caused by a large number of direct calculations is avoided, thereby finally improving the evaluation efficiency and accuracy of the transient stability evaluation of the power system, and ensuring the safe and stable operation of the power system.

[0162] Embodiment 2

[0163] Please refer to Figure 4, which is a structural diagram of a power system transient stability assessment device based on an improved RNN provided by an embodiment of the present invention, comprising: an initial transient stability boundary constraint construction module 201, a first transient stability boundary constraint construction module 202, a fluctuation data acquisition module 203 and a transient stability assessment module 204;

[0164] The initial transient stability boundary constraint building module 201 is used to build the initial transient stability boundary constraint according to the rated operation data of the power system.

[0165] In this embodiment, the initial transient stability boundary constraint construction module 201 includes: an initial transient stability boundary constraint construction unit;

[0166] The rated operation data of the power system used by the initial transient stability boundary constraint construction unit include: the active power of the generator, the reactive power of the generator, the voltage amplitude of the node and the phase angle of the node;

[0167] Constructing a first variable equation group and a second variable equation group of the power system according to the active power of the generator, the reactive power of the generator, the voltage amplitude of the node and the phase angle of the node;

[0168] The power flow equation and the initial state equation of the power system are obtained, and the initial transient stability boundary constraints are constructed by combining the first variable equation group and the second variable equation group.

[0169] The first transient stability boundary constraint building module 202 is used to collect the maximum angle of attack difference data of the generator in the power system, and build the first transient stability boundary constraint in combination with the initial transient stability boundary constraint.

[0170] In this embodiment, the first transient stability boundary constraint construction module 202 includes: a first transient stability boundary constraint construction unit;

[0171] The first transient stability boundary constraint construction unit is used to collect the maximum angle of attack difference data of the generator in the power system according to the preset time window length to obtain the maximum angle of attack difference data;

[0172] According to the maximum angle of attack difference data, a second transient stability boundary constraint is constructed in combination with the initial transient stability boundary constraint; wherein the second transient stability boundary constraint specifically includes:

[0173]

[0174] Where i = 1, 2, 3, ..., N G , N G represents the number of generators in the power system, T represents the length of the simulation time window; δ max The maximum power angle limit to maintain system stability; It represents the absolute power angle value of the i-th generator at time t under the operating state u and the k-th expected fault; Indicates the power angle value in the central inertial coordinate system; T Ji represents the inertia time constant of the i-th generator;

[0175] The second transient stability boundary constraint is converted to obtain a first transient stability boundary constraint.

[0176] In this embodiment, the first transient stability boundary constraint construction unit includes: a first transient stability boundary constraint subunit;

[0177] The first transient stability boundary constraint subunit is used for:

[0178]

[0179] Among them, Φ k (u) Dynamic performance indicators used to represent the power system at a given operating point and anticipated faults; It represents the power angle value in the central inertial coordinate system; It represents the absolute power angle value of the i-th generator at time t under the operating state u and the k-th expected fault. Λ is a binary constant used to distinguish whether the system is stable or not. StableOP means the system is stable, and UnstableOP means the system is unstable.

[0180] The fluctuation data collection module 203 is used to collect the fluctuation data of the power system according to the first transient stability boundary constraint when the power system is disturbed, so as to obtain a fluctuation data set.

[0181] In this embodiment, the fluctuation data collection module 203 includes: a fluctuation data collection unit;

[0182] The fluctuation data acquisition unit is used to obtain a data acquisition direction according to the first transient stability boundary constraint;

[0183] Obtaining a distance between a current operating state of the power system and a transient stable operating state according to the first transient stability boundary constraint;

[0184] The step length coefficient is determined according to the distance between the current operation state of the power system and the transient stable operation state, and then the search step length is obtained;

[0185] When the power system is disturbed, the fluctuation data of the power system is collected according to the data collection direction, the distance between the current operating state of the power system and the transient stable operating state, and the search step size until the power system is in a transient stable operating state, thereby obtaining a fluctuation data set.

[0186] In this embodiment, the fluctuation data acquisition unit includes: a data acquisition direction determination subunit;

[0187] The data acquisition direction determination subunit is used to solve the first transient stability boundary constraint by a first-order derivative to obtain a first-order derivative function of a stability index;

[0188] The data collection direction is obtained according to the first-order derivative function of the stability indicator.

[0189] In this embodiment, the fluctuation data acquisition unit includes: a search step length determination subunit;

[0190] The search step length determination subunit is used to obtain the interval of the distance between the current operating state of the power system and the transient stable operating state according to the preset distance interval set;

[0191] Determine the step size coefficient according to the interval;

[0192] A vector value of the current operating state of the current power system is obtained, and the search step length is determined in combination with the step length coefficient.

[0193] The transient stability assessment module 204 is used to input the fluctuation data into a preset transient stability assessment model to obtain a transient stability assessment result of the power system.

[0194] In this embodiment, the transient stability assessment module 204 includes: a transient stability assessment unit;

[0195] The transient stability assessment unit is used to construct a transient stability assessment model of the power system according to a preset model construction method;

[0196] Inputting the fluctuation data set into the power system transient stability assessment model to obtain a transient stability assessment value;

[0197] According to the preset transient stability assessment threshold, the transient stability assessment result of the power system is obtained in combination with the transient stability assessment value.

[0198] In this embodiment, the transient stability assessment unit includes: a power system transient stability assessment model subunit;

[0199] The power system transient stability assessment model subunit is used for the power system transient stability assessment model and includes: a first layer neural network, a second layer neural network and an output layer;

[0200] The output end of the first neural network is connected to the input end of the second neural network, and the output end of the second neural network is connected to the input end of the output layer; the input end of the first neural network is used to receive the fluctuation data set of the power system transient stability assessment model; the output layer is used to output the transient stability assessment value of the power system transient stability assessment model;

[0201] Wherein, the first layer of neural network is a dynamically stable recursive neural network, which is used to classify the fluctuation data set input into the power system transient stability assessment model;

[0202] The hidden state expression of the dynamically stable recurrent neural network is:

[0203]

[0204] in, for representing the hidden state of the dynamically stable recurrent neural network; for representing an input of the dynamically stable recurrent neural network; is the jump coefficient; is a diagonal matrix with α as the diagonal entry i (j) , is the activation function; the parameter m∈N is used to indicate the number of weighted skip connections across time steps;

[0205] The second layer of neural network is a fully connected layer, which is used to reduce the dimension of the output data of the first layer of neural network until the required dimension of the output layer is reached;

[0206] The output layer is used to classify and analyze the output data of the fully connected layer to obtain a transient stability assessment value.

[0207] This embodiment constructs the initial transient stability boundary constraint through the rated operation data of the power system, and then builds the first transient stability boundary constraint in combination with the maximum angle of attack difference data of the generator in the power system. The rated operation data and the maximum angle of attack difference data of the generator that are closely related to the safe and stable operation of the power system are used to make the transient stability boundary constraint fully adapt to the characteristics of the power system, and the rated operation data and the maximum angle of attack difference data that are conveniently collected simplify the construction process of the transient stability boundary constraint and reduce the calculation complexity, and finally improve the efficiency and accuracy of the transient stability boundary solution. Through the accurate transient stability boundary, the fluctuation data of the power system under large disturbance can be collected, and then the transient stability evaluation result of the power system is obtained in combination with the transient stability evaluation model. By combining the accurate power system fluctuation data with the transient stability evaluation model, compared with the existing time domain simulation and direct method, the characteristics of transient stability in the fluctuation data can be fully and quickly mined, the accuracy of transient stability evaluation is improved, and the inefficiency caused by a large number of direct calculations is avoided, thereby finally improving the evaluation efficiency and accuracy of the transient stability evaluation of the power system, and ensuring the safe and stable operation of the power system.

[0208] In summary, the embodiment of the present invention constructs the initial transient stability boundary constraint through the rated operation data of the power system, and then builds the first transient stability boundary constraint in combination with the maximum angle of attack difference data of the generator in the power system, so that the transient stability boundary constraint can fully adapt to the characteristics of the power system by using the rated operation data and the maximum angle of attack difference data of the generator that are closely related to the safe and stable operation of the power system, and simplifies the construction process of the transient stability boundary constraint and reduces the calculation complexity through the convenient collection of rated operation data and maximum angle of attack difference data, and finally improves the efficiency and accuracy of solving the transient stability boundary. Through the accurate transient stability boundary, the fluctuation data of the power system under large disturbance can be collected, and then the transient stability evaluation result of the power system is obtained by combining the transient stability evaluation model. By combining the accurate power system fluctuation data with the transient stability evaluation model, compared with the existing time domain simulation and direct method, it can fully and quickly mine the characteristics of transient stability in the fluctuation data, improve the accuracy of transient stability evaluation, avoid the inefficiency caused by a large number of direct calculations, and finally improve the evaluation efficiency and accuracy of transient stability evaluation of the power system, and ensure the safe and stable operation of the power system.

[0209] 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 transient stability assessment of power system based on improved RNN, characterized in that: include: Constructing initial transient stability boundary constraints based on rated operation data of the power system; Collecting maximum angle of attack difference data of generators in the power system, and constructing a first transient stability boundary constraint in combination with the initial transient stability boundary constraint; When the power system is disturbed, collecting fluctuation data of the power system according to the first transient stability boundary constraint to obtain a fluctuation data set; A power system transient stability assessment model is constructed, and the fluctuation data set is input into the power system transient stability assessment model to obtain a transient stability assessment result of the power system.

2. A method for evaluating transient stability of a power system based on an improved RNN as claimed in claim 1, characterized in that: The initial transient stability boundary constraint is constructed according to the rated operation data of the power system, specifically including: The rated operation data of the power system include: active power of the generator, reactive power of the generator, voltage amplitude of the node and phase angle of the node; Constructing a first variable equation group and a second variable equation group of the power system according to the active power of the generator, the reactive power of the generator, the voltage amplitude of the node and the phase angle of the node; The power flow equation and the initial state equation of the power system are obtained, and the initial transient stability boundary constraints are constructed by combining the first variable equation group and the second variable equation group.

3. A method for evaluating transient stability of a power system based on an improved RNN as claimed in claim 2, characterized in that: The collecting of the maximum angle of attack difference data of the generator in the power system and constructing the first transient stability boundary constraint in combination with the initial transient stability boundary constraint specifically includes: According to a preset time window length, the maximum angle of attack difference data of the generator in the power system is collected to obtain the maximum angle of attack difference data; According to the maximum angle of attack difference data, a second transient stability boundary constraint is constructed in combination with the initial transient stability boundary constraint; wherein the second transient stability boundary constraint specifically includes: Where i = 1, 2, 3, ..., N G , N G represents the number of generators in the power system, T represents the length of the simulation time window; δ max The maximum power angle limit to maintain system stability; It represents the absolute power angle value of the i-th generator at time t under the operating state u and the k-th expected fault; Indicates the power angle value in the central inertial coordinate system; T Ji represents the inertia time constant of the i-th generator; The second transient stability boundary constraint is converted to obtain a first transient stability boundary constraint.

4. A method for evaluating transient stability of a power system based on an improved RNN as claimed in claim 3, characterized in that: The first transient stability boundary constraint specifically includes: Among them, Φ k (u) is used to represent the dynamic performance index of the power system at a given operating point and expected faults; θ k represents the weight function, which is used to measure whether the system is in a stable state during the fault period, δ k (t|u) represents the dynamic state variables of the system as a function of time when the kth expected fault occurs, and depends on the given static variable u. It represents the power angle value in the central inertial coordinate system; It represents the absolute power angle value of the i-th generator at time t under the operating state u and the k-th expected fault. Λ is a binary constant used to distinguish whether the system is stable or not. StableOP means the system is stable, and UnstableOP means the system is unstable.

5. A method for evaluating transient stability of a power system based on an improved RNN as claimed in claim 1, 3 or 4, characterized in that: When the power system is disturbed, collecting the fluctuation data of the power system according to the first transient stability boundary constraint to obtain the fluctuation data set specifically includes: Obtaining a data acquisition direction according to the first transient stability boundary constraint; Obtaining a distance between a current operating state of the power system and a transient stable operating state according to the first transient stability boundary constraint; The step length coefficient is determined according to the distance between the current operation state of the power system and the transient stable operation state, and then the search step length is obtained; When the power system is disturbed, the fluctuation data of the power system is collected according to the data collection direction, the distance between the current operating state of the power system and the transient stable operating state, and the search step size until the power system is in a transient stable operating state, thereby obtaining a fluctuation data set.

6. A method for evaluating transient stability of a power system based on an improved RNN as claimed in claim 5, characterized in that: The obtaining of the data acquisition direction according to the first transient stability boundary constraint specifically includes: Solving the first-order derivative of the first transient stability boundary constraint to obtain a first-order derivative function of the stability index; The data collection direction is obtained according to the first-order derivative function of the stability indicator.

7. A method for evaluating transient stability of a power system based on an improved RNN as claimed in claim 5, characterized in that: The step length coefficient is determined according to the distance between the current operation state of the power system and the transient stable operation state, and then the search step length is obtained, specifically including: Obtaining the interval of the distance between the current operating state of the power system and the transient stable operating state according to the preset distance interval set; Determine the step size coefficient according to the interval; A vector value of the current operating state of the current power system is obtained, and the search step length is determined in combination with the step length coefficient.

8. The method for evaluating transient stability of a power system based on an improved RNN according to claim 1, characterized in that: The step of constructing a transient stability assessment model for a power system, inputting the fluctuation data set into the transient stability assessment model for a power system, and obtaining a transient stability assessment result for a power system specifically includes: Construct a power system transient stability assessment model according to a preset model construction method; Inputting the fluctuation data set into the power system transient stability assessment model to obtain a transient stability assessment value; According to the preset transient stability assessment threshold, the transient stability assessment result of the power system is obtained in combination with the transient stability assessment value.

9. A method for evaluating transient stability of a power system based on an improved RNN as claimed in claim 8, characterized in that: The power system transient stability assessment model specifically includes: The power system transient stability assessment model comprises: a first layer neural network, a second layer neural network and an output layer; The output end of the first neural network is connected to the input end of the second neural network, and the output end of the second neural network is connected to the input end of the output layer; the input end of the first neural network is used to receive the fluctuation data set of the power system transient stability assessment model; the output layer is used to output the transient stability assessment value of the power system transient stability assessment model; Wherein, the first layer of neural network is a dynamically stable recursive neural network, which is used to classify the fluctuation data set input into the power system transient stability assessment model; The hidden state expression of the dynamically stable recurrent neural network is: in, for representing the hidden state of the dynamically stable recurrent neural network; for representing an input of the dynamically stable recurrent neural network; is the jump coefficient; is a diagonal matrix with diagonal entries of is the activation function; the parameter m∈N is used to indicate the number of weighted skip connections across time steps; The second layer of neural network is a fully connected layer, which is used to reduce the dimension of the output data of the first layer of neural network until the required dimension of the output layer is reached; The output layer is used to classify and analyze the output data of the fully connected layer to obtain a transient stability assessment value.

10. A power system transient stability assessment device based on improved RNN, characterized in that: include: An initial transient stability boundary constraint building module, a first transient stability boundary constraint building module, a fluctuation data acquisition module and a transient stability assessment module; Wherein, the initial transient stability boundary constraint construction module is used to construct the initial transient stability boundary constraint according to the rated operation data of the power system; The first transient stability boundary constraint building module is used to collect the maximum angle of attack difference data of the generator in the power system, and to build the first transient stability boundary constraint in combination with the initial transient stability boundary constraint; The fluctuation data collection module is used to collect the fluctuation data of the power system according to the first transient stability boundary constraint when the power system is disturbed, so as to obtain a fluctuation data set; The transient stability assessment module is used to construct a transient stability assessment model for the power system, input the fluctuation data set into the transient stability assessment model for the power system, and obtain a transient stability assessment result for the power system.

Images