Power grid voltage safety and stability control method and system based on TCRN and GOA

Abstract

The invention discloses a power grid voltage safety and stability control method and system based on TCRN and GOA, and relates to the technical field of power grid voltage safety and stability control, and the method comprises the steps: giving a voltage reduction interval weight to refine the power grid stability influence, and building a voltage stability evaluation quantitative index; constructing a voltage stability prediction model based on the voltage stability evaluation quantitative index, and predicting the relation between the voltage state and the stability; and through a prediction result, constructing a voltage stability control model, and outputting an optimal generator shedding and load shedding strategy. According to the method, a voltage stability prediction model based on a time sequence convolution residual network is constructed, the relation between sensitive node voltage time sequence characteristics and voltage stability is mapped, a control strategy is output by constructing a voltage stability control model, the pelican optimization algorithm is used for solving the control model with the advantages of being high in convergence speed and high in search capability, and the voltage stability prediction model is used for predicting the voltage stability. And outputting the optimal generator shedding and load shedding strategy.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid voltage safety and stability control, and in particular to a power grid voltage safety and stability control method and system based on TCRN and GOA. Background Art

[0002] The safety and stability control device of the power grid is a secondary device used to ensure that the power grid maintains system stability when it is disturbed. It is usually installed in key plants and stations such as power plants and substations. Multiple safety and stability control devices are jointly controlled through communication to form a safety and stability control system. With the access of a high proportion of distributed renewable energy to the power grid, the complexity and uncertainty of system operation have increased significantly. Compared with traditional synchronous generators, distributed renewable energy has weak anti-interference ability and low mechanical rotational inertia, which seriously weakens its ability to regulate grid voltage. When a large interference fault occurs in the system, the complex power flow distribution may cause grid instability and even cause the system to be decoupled. Therefore, it is urgent to study safer, more reliable, flexible and accurate safety and stability control methods to improve the stability and reliability of power grid operation.

[0003] At present, existing studies have sorted out the fault types of safety and stability control systems, but the fault diagnosis method has not yet been fully applied to the system. When facing large-scale power grid data, relying on manual extraction of fault features is not only inefficient, but also difficult to accurately predict the stability of the power grid. At this stage, relatively mature, reliable and low-cost power grid safety control systems are mainly implemented in the form of strategy tables. When a large interference fault occurs, the safety and stability control device searches for strategies and implements countermeasures based on the fault type and operating status. This method has the advantage of fast response. However, in complex power grids with large-scale access to distributed power sources, this method is difficult to fully cover all fault types, and is prone to decision-making errors, which cannot meet the needs of safe operation of the power grid.

[0004] To sum up, in order to cope with the problems of difficulty in accurately judging the fault type and decision-making errors caused by the large-scale access of distributed power sources, research on safer, more reliable, flexible and accurate grid voltage safety and stability control strategies has become an important problem and research focus of current power grid operation. Summary of the invention

[0005] In view of the above-mentioned problems, the present invention is proposed.

[0006] Therefore, the technical problem solved by the present invention is: the existing power grid safety and stability control method has the problems of difficult to accurately judge the fault type, lack of flexibility in decision-making methods, insufficient data processing capabilities, and how to achieve a safer, more reliable, flexible and accurate power grid voltage safety and stability control strategy.

[0007] To solve the above technical problems, the present invention provides the following technical solutions: a method for safe and stable control of power grid voltage based on TCRN and GOA, comprising refining the impact of power grid stability by assigning voltage reduction interval weights, and establishing a quantitative indicator for voltage stability evaluation; based on the quantitative indicator for voltage stability evaluation, constructing a voltage stability prediction model to predict the relationship between voltage state and stability; through the prediction results, constructing a voltage stability control model to output the optimal machine cutting and load cutting strategy.

[0008] As a preferred solution of the grid voltage safety and stability control method based on TCRN and GOA described in the present invention, wherein: the assigning of voltage reduction interval weights to refine the grid stability impact includes collecting key node voltage time series data during grid operation, allocating dynamic weights to quantify the impact of each interval on grid stability according to different voltage reduction intervals, and establishing a binary table. cr , τ cr >,u cr is the voltage threshold value, which indicates the critical value of voltage under stable conditions. When the bus voltage is lower than u cr When τ cr is the time threshold value, indicating that the voltage drops to the critical voltage value u cr The following time threshold; if the voltage is lower than u cr The time exceeds the time threshold τ cr , the power system enters an unstable state; if the voltage reduction value satisfies multiple binary tables [ cr1 , τ cr1 >, cr2 , τ cr2 >, …, crn , τ crn >], the power system is stable.

[0009] As a preferred solution of the grid voltage safety and stability control method based on TCRN and GOA described in the present invention, wherein: the establishment of the voltage stability evaluation quantitative index includes refining the grid stability impact by assigning a voltage reduction interval weight, calculating the voltage stability evaluation quantitative index ξ m , expressed as:

[0010]

[0011] Among them, K r , K n is the weight, u N is the rated voltage, n is the total number of binary meters, u(t) is the value of the bus voltage at time t, t′ r ,t r The voltage is higher and lower than u cr The moment, t n is the time when the voltage starts to recover, t′​​​​n The voltage is fully restored to the rated voltage u N time; if the power grid is critically stable, then ξ m The value is taken as 0, and all weights are solved one by one, expressed as:

[0012]

[0013] When the minimum value of transient voltage drop is greater than u in the binary table cr When , the voltage stability evaluation quantitative index is calculated, which is expressed as:

[0014]

[0015] Among them, τ crr 、u crr are the time threshold and voltage threshold of the rth binary table, t 0 The voltage drops moment.

[0016] As a preferred solution of the grid voltage safety and stability control method based on TCRN and GOA described in the present invention, wherein: the construction of the voltage stability prediction model includes constructing a voltage stability prediction model based on a time series convolutional residual network based on the voltage stability evaluation quantitative index, and extracting features from the bus key node voltage sequence, which is expressed as:

[0017]

[0018] in, are the network input and output of the Lth layer, con is the result of multiple convolutions combined along the feature axis, f(·) is the activation function, K is the convolution kernel parameter, b is the bias, and s is the distance that the convolution moves along the time axis at a specific step size. is the temporal convolution matrix.

[0019] As a preferred solution of the method for controlling the safe and stable voltage of a power grid based on TCRN and GOA of the present invention, the construction of the voltage stability prediction model further includes extracting features by short-circuiting, which is expressed as:

[0020] H(x)=x+F(x)

[0021] Among them, F(x) is the output feature of input x after weighting, and H(x) is the output feature using short-circuit wires; the activation unit is modified and expressed as:

[0022] ReLU(x in )=max(0,x in )

[0023] Among them, ReLU is the rectified linear activation function, xin is the input of the corrected linear activation unit; the network training efficiency is optimized through the batch normalization layer, expressed as:

[0024]

[0025] Among them, x is the network adaptive change input, is the normalized network adaptive change output, m is the number of samples, γ and β are the scaling coefficient and translation coefficient of the network respectively, and ε is a positive number to prevent the denominator from being zero.

[0026] As a preferred solution of the grid voltage safety and stability control method based on TCRN and GOA described in the present invention, wherein: the construction of the voltage stability control model includes constructing a voltage stability control model based on the Pelican optimization algorithm, taking the machine cutting and load cutting capacity as the optimization target, and constructing a control model with the minimum machine cutting and load cutting capacity, which is expressed as:

[0027]

[0028] Among them, ΔP G,i is the generator removal capacity of the i-th node, ΔP L,j u j is the load shedding capacity of the jth node, N G 、N L are the number of generator and load shedding nodes, u i 、u j are generator and load shedding coefficients respectively; load shedding coefficient u j is a continuous variable between 0 and 1, and the generator removal coefficient u is calculated. i , expressed as:

[0029]

[0030] Setting voltage stability evaluation index ξ m Above the threshold ξ, the constraint formula is expressed as:

[0031]

[0032] Among them, ξ k,0 is the voltage stability evaluation index of the kth node when no stability control measures are implemented after a large interference fault, λ k,i is the voltage stability control sensitivity coefficient, λ k,j is the load adjustment sensitivity coefficient; the voltage stability control sensitivity coefficient is calculated and expressed as:

[0033]

[0034] Among them, ξ kis the adjusted voltage stability evaluation index value, ΔP i is the power removal of the i-th generator; the power balance constraint is defined as:

[0035]

[0036] Among them, S b,min , S b,max They are respectively the minimum and maximum values ​​of the power unbalance capacity when the generator frequency regulation factor is introduced.

[0037] As a preferred solution of the grid voltage safety and stability control method based on TCRN and GOA described in the present invention, wherein: the output optimal machine cutting and load cutting strategy includes setting the machine cutting and load cutting capacity to an individual of the Pelican optimization algorithm, initializing the Pelican optimization algorithm parameters including the Pelican position, speed, number of iterations, and population size, and adjusting the machine cutting and load cutting capacity according to the Pelican position; according to the system voltage node bus sequence, the voltage stability prediction based on the time series convolution residual network is obtained, if the voltage is stable, the voltage stability control model objective function is calculated, the optimal position of the Pelican is determined according to the objective function, and the Pelican position is updated; if the voltage is unstable, the Pelican position is not updated; when all Pelican positions are updated, if they meet the convergence or reach the maximum number of iterations, the final machine cutting and load cutting capacity is output; if they do not meet the convergence or the maximum number of iterations is not reached, the machine cutting and load cutting capacity is re-adjusted according to the Pelican position.

[0038] Another object of the present invention is to provide a power grid voltage safety and stability control system based on TCRN and GOA, which can construct a voltage stability prediction model based on voltage stability evaluation quantitative indicators, predict the relationship between voltage state and stability, and solve the problem of inaccurate fault type judgment in current power grid safety and stability control technology.

[0039] As a preferred solution of the power grid voltage safety and stability control system based on TCRN and GOA described in the present invention, it includes: a voltage stability assessment module, a voltage stability prediction module, and a voltage stability control optimization module; the voltage stability assessment module is used to refine the power grid stability impact by assigning voltage reduction interval weights and establish voltage stability assessment quantitative indicators; the voltage stability prediction module is used to construct a voltage stability prediction model based on the voltage stability assessment quantitative indicators to predict the relationship between voltage state and stability; the voltage stability control optimization module is used to construct a voltage stability control model through the prediction results and output the optimal machine cutting and load cutting strategy.

[0040] A computer device includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement a step of a method for controlling a power grid voltage safety and stability based on TCRN and GOA.

[0041] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a method for controlling a power grid voltage safety and stability based on TCRN and GOA.

[0042] Beneficial effects of the present invention: The grid voltage safety and stability control method based on TCRN and GOA provided by the present invention constructs a voltage stability prediction model based on a timing convolution residual network by utilizing the advantages of less information loss, wide receptive field and strong deep feature extraction capability of the residual network, maps the relationship between the voltage timing characteristics of sensitive nodes and voltage stability, outputs the control strategy by constructing a voltage stability control model, and solves the control model by utilizing the advantages of fast convergence speed and strong search capability of the Pelican optimization algorithm to obtain the optimal machine cutting and load cutting action measures. The grid voltage safety and stability control method based on the timing convolution residual network and the Pelican optimization algorithm improves the accuracy of grid voltage safety and stability prediction, and improves the safe and stable operation level after a grid fault. The present invention achieves better results in terms of accuracy, safety and reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0044] Figure 1 An overall flow chart of a grid voltage safety and stability control method based on TCRN and GOA provided for the first embodiment of the present invention.

[0045] Figure 2 A timing convolutional residual network component diagram of a grid voltage safety and stability control method based on TCRN and GOA provided in the first embodiment of the present invention.

[0046] Figure 3 A voltage stability control flow chart of a grid voltage safety and stability control method based on TCRN and GOA is provided for the first embodiment of the present invention.

[0047] Figure 4 A power system geographical wiring diagram of a power grid voltage safety and stability control method based on TCRN and GOA is provided for the second embodiment of the present invention.

[0048] Figure 5 An equivalent circuit diagram of a grid voltage safety and stability control method based on TCRN and GOA provided in the second embodiment of the present invention.

[0049] Figure 6 A relationship diagram between the voltage stability prediction accuracy and the number of neurons of a grid voltage security and stability control method based on TCRN and GOA provided in the second embodiment of the present invention.

[0050] Figure 7 A relationship diagram between the loss function and the number of iterations of a grid voltage security and stability control method based on TCRN and GOA provided in the second embodiment of the present invention.

[0051] Figure 8 A comparison diagram of voltage stability prediction results of different methods of a grid voltage security and stability control method based on TCRN and GOA provided in the second embodiment of the present invention.

[0052] Fig. 9 A bus voltage recovery effect diagram of a grid voltage safety and stability control method based on TCRN and GOA provided in the second embodiment of the present invention.

[0053] Fig.10 A comparison diagram of objective function results of different methods of a grid voltage security and stability control method based on TCRN and GOA provided in the second embodiment of the present invention.

[0054] Fig.11 A module diagram of a grid voltage safety and stability control system based on TCRN and GOA is provided for the third embodiment of the present invention. DETAILED DESCRIPTION

[0055] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0056] Example 1, reference Figure 1-Figure 3 , which is an embodiment of the present invention, provides a grid voltage safety and stability control method based on TCRN and GOA, including:

[0057] Among them, TCRN (Temporal Convolutional Residual Network) is the temporal convolutional residual network, and GOA (Gannet Optimizer Algorithm) is the Pelican optimization algorithm.

[0058] S1: The impact on grid stability is refined by assigning weights to the voltage reduction intervals, and quantitative indicators for voltage stability assessment are established.

[0059] Furthermore, the voltage reduction interval weights are assigned to refine the impact on grid stability, including collecting voltage time series data of key nodes during grid operation, assigning dynamic weights to quantify the impact of each interval on grid stability according to different voltage reduction intervals, and establishing a binary table. cr , τ cr ),u cr is the voltage threshold value, which indicates the critical value of voltage under stable conditions. When the bus voltage is lower than u cr When τ cr is the time threshold value, indicating that the voltage drops to the critical voltage value u cr The following time threshold; if the voltage is lower than u cr The time exceeds the time threshold τ cr , the power system enters an unstable state; if the voltage reduction value satisfies multiple binary tables [ cr1 , τ cr1 >, cr2 , τ cr2 >, …, crn , τ crn >], the power system is stable.

[0060] It should be noted that establishing the voltage stability evaluation quantitative index includes refining the grid stability impact by assigning voltage reduction interval weights and calculating the voltage stability evaluation quantitative index ξ m , expressed as:

[0061]

[0062] Among them, K r , K n is the weight, u N is the rated voltage, n is the total number of binary meters, u(t) is the value of the bus voltage at time t, t′ r ,t r The voltage is higher and lower than u cr The moment, t n is the time when the voltage starts to recover, t′ n The voltage is fully restored to the rated voltage u N time; if the power grid is critically stable, then ξ m The value is taken as 0, and all weights are solved one by one, expressed as:

[0063]

[0064] When the minimum value of transient voltage drop is greater than u in the binary table cr When , the voltage stability evaluation quantitative index is calculated, which is expressed as:

[0065] ​​​​

[0066] Among them, τ crr 、u crr are the time threshold and voltage threshold of the rth binary table, t 0 The voltage drops moment.

[0067] It should also be noted that by assigning dynamic weights to different voltage reduction intervals, the actual impact of each interval on the stability of the power grid can be analyzed in detail. Compared with the traditional fixed weights or simple average distribution method, the dynamic weight distribution is closer to the actual operating state and can capture the complex dynamic behavior of voltage fluctuations. The voltage threshold value represents the critical voltage under stable conditions, and the time threshold value defines the time range allowed when the voltage is lower than the critical value. The dynamic characteristics of voltage and time are combined through a binary table, and the time constraint of voltage fluctuations is established. The dynamic conditions of the power grid from a stable state to an unstable state are clarified, and the impact of normal fluctuations and severe instability are distinguished. The response range can be reduced to reduce the risk of misjudgment and avoid erroneous control actions caused by short-term fluctuations. At the same time, it can ensure timely identification and response in critical states. By combining the dynamic weights of different voltage reduction intervals with time series data, the comprehensive impact of voltage fluctuations on power grid stability can be quantified. Especially in the case of superposition of multi-interval fluctuations, the global stability index is calculated through the combination of binary table and weights, which provides a unified and dynamic voltage stability evaluation method, solves the defect that traditional methods are difficult to deal with multi-interval dynamic fluctuations, accurately calculates the global stability of the power grid, and provides a quantifiable indicator basis for subsequent voltage prediction and control.

[0068] S2: Based on the voltage stability assessment quantitative indicators, a voltage stability prediction model is constructed to predict the relationship between voltage status and stability.

[0069] Furthermore, the construction of the voltage stability prediction model includes constructing a voltage stability prediction model based on the time series convolution residual network based on the voltage stability evaluation quantitative index, and extracting the features of the voltage sequence of the key bus nodes, which is expressed as:

[0070]

[0071] in, are the network input and output of the Lth layer, con is the result of multiple convolutions combined along the feature axis, f(·) is the activation function, K is the convolution kernel parameter, b is the bias, and s is the distance that the convolution moves along the time axis at a specific step size. is the temporal convolution matrix.

[0072] A voltage stability prediction model based on the time series convolution residual network is constructed. The width can be increased or reduced by lateral convolution to adapt to the number of channels of the input features. All channels j are covered by the receptive field, reducing the lateral movement calculation of the convolution kernel, greatly saving the calculation time and improving the network training efficiency. The longitudinal convolution is performed downward along the direction of the movement of the time axis i according to the node voltage time series characteristics, and the information of the feature sequence time axis direction is mined. A new feature matrix is ​​formed after each sliding window calculation.

[0073] It should be noted that building a voltage stability prediction model also includes adopting Figure 2 The short-circuit wiring method shown in the figure is used for feature extraction, which is expressed as:

[0074] H(x)=x+F(x)

[0075] Among them, F(x) is the output feature of the input x after weighting, and H(x) is the output feature using short-circuit wiring. Introducing short connections between internal inputs and outputs can form layer backoff. By short-circuiting, the internal weight changes from H(x) to H(x)-x. The gradient can quickly and completely approach the input layer during the back propagation process, thereby improving the parameter optimization capability.

[0076] The activation unit is modified and expressed as:

[0077] ReLU(x in )=max(0,x in )

[0078] Among them, ReLU is the rectified linear activation function, x in To correct the input of the linear activation unit; by correcting the linear activation unit, the nonlinear learning ability of the model can be improved, and the gradient dissipation phenomenon of complex data in the network model feature learning can be prevented.

[0079] The network training efficiency is optimized through the batch normalization layer, which is expressed as:

[0080]

[0081] Among them, x is the network adaptive change input, is the normalized network adaptive change output, m is the number of samples, γ and β are the scaling factor and translation factor of the network respectively, and ε is a positive number to prevent the denominator from being zero. The batch normalization layer unifies all data scales to reduce the hyperparameter sensitivity of the network, thereby optimizing network performance and improving training speed.

[0082] It should also be noted that the voltage stability prediction model constructed in the present invention is based on the time series convolutional residual network, and the model solving process is as follows:

[0083] Step 1: Data preprocessing. The voltage time series data of the key node bus is used as the network input to evaluate the voltage stability level. In order to improve the convergence speed and prediction accuracy of the model, the input data is normalized so that the data range is unified within the specified interval to reduce the error accumulation and calculation complexity caused by data scale differences.

[0084] Step 2: Determine the model structure of the temporal convolutional residual network. Determine the model structure of the temporal convolutional residual network, including the number of network layers and the number of neurons in each layer. This model uses the ReLU activation function, which is characterized by being able to make the output of some neurons zero during model training, thereby introducing sparsity, effectively reducing the complexity of the model, and preventing the occurrence of overfitting problems. At the same time, the introduction of the residual structure enhances the network's ability to extract deep features and avoids the gradient vanishing problem in deep networks.

[0085] Step 3: Determine network parameters with training set. Input the bus voltage time series data of key nodes in the training set into the network, and forward propagate through time series convolution and residual structure to output voltage stability evaluation quantitative index ξ m , the evaluation results are compared with the threshold:

[0086] If the indicator ξ m If it is higher than the threshold ξ, it means the system voltage is stable;

[0087] If the indicator ξ m If it is lower than the threshold ξ, it means the system voltage is unstable;

[0088] According to the error between the predicted output and the actual voltage state, a loss function is constructed, and the network parameters are updated through back propagation according to the principle of minimizing the loss. After multiple iterative optimizations, an accurate and reliable temporal convolutional residual network model is finally trained for voltage stability prediction.

[0089] It should also be noted that the voltage stability prediction model based on the time series convolution residual network (TCRN) constructed by the present invention fully utilizes the advantages of time series convolution in less information loss, wide receptive field and strong deep feature extraction capability of the residual network. By using time series convolution instead of traditional convolution operation, the model significantly expands the receptive field (i.e. the ability of the model to perceive the input data range) when convolving along the time axis, and includes all lateral features at the same time, enhancing the mining capability of voltage time series information and lateral features. In addition, the residual network structure further improves the extraction efficiency of deep features and avoids the gradient vanishing problem in deep networks.

[0090] S3: Based on the prediction results, a voltage stability control model is constructed to output the optimal generator and load shedding strategies.

[0091] Furthermore, the construction of the voltage stability control model includes constructing a voltage stability control model based on the Pelican optimization algorithm, taking the machine cutting and load cutting capacity as the optimization target, and constructing a control model with the minimum machine cutting and load cutting capacity, which is expressed as:

[0092]

[0093] Where ΔP G，i is the generator removal capacity of the i-th node, ΔP L，j u j is the load shedding capacity of the jth node, N G 、N L are the number of generator and load shedding nodes, u i 、u j are generator and load shedding coefficients respectively; load shedding coefficient u j is a continuous variable between 0 and 1, and the generator removal coefficient u is calculated. i , expressed as:

[0094]

[0095] Among them, the load shedding factor u j It is determined by considering the power supply reliability under the premise of maintaining voltage stability. When a large interference fault occurs, the load is cut off by a certain proportion, so u j is a continuous variable between 0 and 1, and the generator is usually shut down as a whole, so the generator removal coefficient u i is a discrete integer.

[0096] Setting voltage stability evaluation index ξ m Above the threshold ξ, the constraint formula is expressed as:

[0097]

[0098] Among them, ξ k,0 is the voltage stability evaluation index of the kth node when no stability control measures are implemented after a large interference fault, λ k,i is the voltage stability control sensitivity coefficient, λ k,j is the load adjustment sensitivity coefficient; the voltage stability control sensitivity coefficient is calculated and expressed as:

[0099]

[0100] Among them, ξ k is the adjusted voltage stability evaluation index value, ΔP i is the power removal of the i-th generator; the power balance constraint is defined as:

[0101]

[0102] Among them, S b,min , S b,max They are respectively the minimum and maximum values ​​of the power unbalance capacity when the generator frequency regulation factor is introduced.

[0103] It should be noted that outputting the optimal machine cutting and load shedding strategy includes setting the machine cutting and load shedding capacity to the individual of the Pelican optimization algorithm, initializing the parameters of the Pelican optimization algorithm including the pelican position, speed, number of iterations, and population size, and adjusting the machine cutting and load shedding capacity according to the pelican position; according to the system voltage node bus sequence, the voltage stability prediction based on the time series convolution residual network is obtained. If the voltage is stable, the voltage stability control model objective function is calculated, the optimal pelican position is determined according to the objective function, and the pelican position is updated; if the voltage is unstable, the pelican position is not updated; when all pelican positions are updated, if they meet the convergence or reach the maximum number of iterations, the final machine cutting and load shedding capacity is output; if they do not meet the convergence or the maximum number of iterations is not reached, the machine cutting and load shedding capacity is re-adjusted according to the pelican position.

[0104] It should also be noted that the Pelican Optimization Algorithm is a new meta-heuristic algorithm that solves the optimal solution through a mathematical model established by the pelican hunting behavior. It has the advantages of fast convergence speed, simple parameters, and strong global and local search capabilities. The Pelican Optimization Algorithm includes three stages: initialization, movement towards the prey, and skimming the water surface.

[0105] In the initialization phase, if the spatial dimension is m and the number of pelicans is n, the position of the i-th pelican in the m-dimensional space is expressed as:

[0106] X i =[X i1 ,X i2 ,…,X im ]

[0107] X i is the position of the i-th pelican, X im is the position of the i-th pelican in m-dimensional space

[0108] All n pelican positions in m-dimensional space are represented as:

[0109]

[0110] Calculate the initial pelican position, expressed as:

[0111] x ij = l j +α·(u j -l j ), i=1,2,…,n; j=1,2,…,m

[0112] Among them, x ij is the j-th dimensional position of the i-th pelican, m is the model solution dimension, l j Solve the minimum value for the j-th dimension model, u j Find the maximum value for the j-th dimension model, α is a random number, and the value of α varies between 0 and 1.

[0113] In the stage of moving towards the prey, the pelican will move towards the prey when identifying the prey's position at this stage. The random distribution of the search space improves the global search ability of the algorithm. The position of the pelican in each iteration is expressed as:

[0114]

[0115] Among them, F p is the prey target value, is the j-th dimension position of the ith pelican after the first stage, σ is a random number, σ varies between 0 and 1, I is an integer, I takes the value of 1 or 2, P j is the prey position in the jth dimension.

[0116] If the target value is optimized, the position update formula is expressed as:

[0117]

[0118] in, is the new objective function of the i-th pelican after the first stage, is the new i-th pelican position.

[0119] In the stage of skimming the water surface, the pelican spreads its wings and puts the prey in its throat after passing the water surface, which improves the local search ability. The pelican position update formula for each iteration is expressed as:

[0120]

[0121] in, is the j-th dimensional position of the ith pelican after the second stage, β is a random number ranging from 0 to 1, R is an integer ranging from 1 to 2, t is the current number of iterations, and T is the maximum number of iterations set.

[0122] If the target value is optimized, the position update formula is expressed as:

[0123]

[0124] in, is the new objective function of the i-th pelican after the second stage, is the new i-th pelican position.

[0125] It should also be noted that the load shedding capacity is set as the individual of the Pelican optimization algorithm, and the voltage stability control process is as follows Figure 3 As shown, the specific steps of the solution process are as follows:

[0126] A1. Initialize the pelican optimization algorithm parameters including pelican position, speed, number of iterations, and population size.

[0127] A2. Adjust the load cutting capacity of the cutting machine according to the position of each pelican.

[0128] A3. According to the system voltage node bus sequence, the voltage stability prediction based on the time series convolution residual network is obtained. If the voltage is stable, the objective function value of the voltage stability control model is calculated, and the optimal position of the pelican is found according to the objective function, and the pelican position is updated; if the voltage is unstable, the pelican position is not updated.

[0129] A4. Update the positions and speeds of all pelicans and determine whether they have converged or reached the maximum number of iterations. If not, return to step A2. Otherwise, output the final load shedding capacity of each machine.

[0130] Example 2, reference Figure 4-Figure 10 , which is an embodiment of the present invention, provides a grid voltage safety and stability control method based on TCRN and GOA. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0131] In order to verify the effectiveness of the solution of the present invention, a regional power system model was built using PSASP simulation software for transient calculation. The geographical wiring diagram is shown in Figure 4 As shown, including busbar, power plant, converter station, the equivalent circuit is as follows Figure 5 As shown, the total number of nodes is 85, the rated power of the entire system is 10000MW, and the rated voltage is 800kV.

[0132] First, the input of the voltage stability prediction model network based on the time series convolutional residual network is the key node voltage time series: 35 node active bus voltage time series values, and the number of sampling points is 2000; the output of the network is the voltage stability prediction result, that is, the voltage stability evaluation quantitative index, so the network input data dimension is 2000×35, and the output dimension is 1×1.

[0133] The evaluation indicators of the model mainly include accuracy, missed detection rate and false detection rate. The calculation formula is expressed as:

[0134]

[0135] Among them, A CC is the accuracy, L A is the missed detection rate, F Ais the false positive rate, T is the number of samples predicted correctly, F is the number of samples predicted incorrectly, s is the number of samples predicted stably, and us is the number of samples predicted unstablely.

[0136] Next, determine the number of neurons in the temporal convolutional residual network and set the number of neurons in the temporal convolutional residual network from 100 to 500 respectively; use 600 sets of data sets for training and observe the accuracy A CC Changes to A CC The number of neurons is determined based on the maximum principle. Through experimental results, we get Figure 6 The relationship between the accuracy and the number of neurons in the temporal convolutional residual network is shown in Figure 6 It can be seen that when the number of neurons is 400, A CC Therefore, the number of neurons in the temporal convolutional residual network is set to 400.

[0137] In order to verify the training effect of the network, the relationship between the network loss function and the number of iterations is as follows Figure 7 As shown, from Figure 7 It can be seen that when the number of iterations reaches 500, the network tends to be stable and reaches the minimum loss function value of 0.38, indicating that the network training effect is good.

[0138] In order to verify the superiority of the voltage stability prediction effect of the method of the present invention, the temporal convolutional residual network of the present invention is compared with the convolutional neural network and the residual network, and the prediction results of different methods are obtained as follows: Figure 8 As shown in Table 1, Figure 8 It can be seen from Table 1 that when the number of iterations reaches 500, the voltage stability prediction accuracy of the three methods reaches the highest, and the method of the present invention has the highest accuracy, while the missed alarm rate and false alarm rate are the lowest, which shows that the method of the present invention can well predict the voltage stability of the power system.

[0139] Table 1 Comparison of voltage stability prediction results of different methods

[0140]

[0141] Taking the generator demand and load importance into consideration, some generator nodes and load nodes are used as control nodes, such as Figure 4 The busbars 1, 2, 3, 4, 5, and 12 shown are used as machine cutting points, and the busbars 1, 3, 9, 10, 14, and 15 are used as load cutting points. The cutting capacity of each machine cutting and load cutting point is shown in Table 2 and Table 3.

[0142] Table 2 Capacity that can be removed at each cutting point

[0143]

[0144] Table 3 Capacity that can be removed at each load shedding point

[0145]

[0146] In order to verify the effectiveness of the voltage stability control method of the Pelican optimization algorithm of the present invention, the voltage optimization control measures obtained are shown in Tables 4 and 5.

[0147] Table 4 Voltage stability control measures - capacity cut off at each cutting point

[0148]

[0149]

[0150] Table 5 Voltage stability control measures - capacity that can be removed at each disconnection point

[0151]

[0152] In order to verify the superiority of the voltage stability control of the Pelican optimization algorithm of the present invention, the method of the present invention is compared with the particle swarm algorithm and the genetic algorithm. The bus voltage recovery results of node 9 under different methods are shown in the following figure. Fig. 9 As shown, from Fig. 9 It can be seen that the voltage recovery effect of the method of the present invention is the best, which is close to the voltage before the fault, and the voltage recovery effect of the genetic algorithm is the worst, which shows that the voltage stability control method of the Pelican optimization algorithm of the present invention is better.

[0153] The objective functions of different methods change with the number of iterations as follows: Fig.10 As shown in FIG10 , it can be seen that compared with the particle swarm algorithm and the genetic algorithm, the method of the present invention has the best objective function optimization performance and is more applicable to the power system. Therefore, the present invention is creative.

[0154] Example 3, reference Fig.11 , which is an embodiment of the present invention, provides a power grid voltage safety and stability control system based on TCRN and GOA, including a voltage stability evaluation module, a voltage stability prediction module, and a voltage stability control optimization module.

[0155] The voltage stability assessment module is used to refine the impact of grid stability by assigning voltage reduction interval weights and establish quantitative indicators for voltage stability assessment; the voltage stability prediction module is used to construct a voltage stability prediction model based on the voltage stability assessment quantitative indicators to predict the relationship between voltage state and stability; the voltage stability control optimization module is used to construct a voltage stability control model based on the prediction results and output the optimal machine cutting and load cutting strategies.

[0156] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0157] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0158] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering, or processing in another suitable manner as necessary, and then stored in a computer memory.

[0159] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc. It should be noted that the above embodiments are only used to illustrate the technical solution of the present invention and are not limited. Although the present invention is described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention, which should be included in the scope of the claims of the present invention.

[0160] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for controlling power grid voltage security and stability based on TCRN and GOA, characterized in that: include: By assigning weights to the voltage reduction intervals, the impact on grid stability is refined and quantitative indicators for voltage stability assessment are established. Based on the voltage stability evaluation quantitative indicators, a voltage stability prediction model is constructed to predict the relationship between voltage status and stability; Based on the prediction results, a voltage stability control model is constructed to output the optimal generator and load shedding strategies.

2. The method for controlling the safe and stable voltage of a power grid based on TCRN and GOA according to claim 1, characterized in that: The method of assigning voltage reduction interval weights to refine the impact on grid stability includes collecting voltage time series data of key nodes during grid operation, assigning dynamic weights to quantify the impact of each interval on grid stability according to different voltage reduction intervals, and establishing a binary table. cr , τ cr >,u cr is the voltage threshold value, which indicates the critical value of voltage under stable conditions. When the bus voltage is lower than u cr When τ cr is the time threshold value, indicating that the voltage drops to the critical voltage value u cr The following time thresholds;​ If the voltage is lower than u cr The time exceeds the time threshold τ cr , the power system enters an unstable state; If the voltage reduction value satisfies multiple binary tables [ cr1 , τ cr1 >, cr2 , τ cr2 >, …, crn , τ crn >], the power system is stable.​​​ 3. The method for controlling the safe and stable voltage of a power grid based on TCRN and GOA according to claim 2, characterized in that: The establishment of the voltage stability evaluation quantitative index includes refining the grid stability impact by assigning a voltage reduction interval weight, and calculating the voltage stability evaluation quantitative index ξ m , expressed as: Among them, K r , K n is the weight, u N is the rated voltage, n is the total number of binary meters, u(t) is the value of the bus voltage at time t, t′ r ,t r The voltage is higher and lower than u cr The moment, t n is the time when the voltage starts to recover, t′ n The voltage is fully restored to the rated voltage u N time; If the power grid is critically stable, then m The value is taken as 0, and all weights are solved one by one, expressed as: When the minimum value of transient voltage drop is greater than u in the binary table cr When , the voltage stability evaluation quantitative index is calculated, which is expressed as: Among them, τ crr 、u crr are the time threshold value and voltage threshold value of the rth binary table respectively, and t0 is the moment when the voltage decreases.

4. The method for controlling the safe and stable voltage of a power grid based on TCRN and GOA according to claim 3, characterized in that: The construction of the voltage stability prediction model includes constructing a voltage stability prediction model based on a time series convolution residual network based on the voltage stability evaluation quantitative index, and extracting features of the busbar key node voltage sequence, which is expressed as: in, are the network input and output of the Lth layer, con is the result of multiple convolutions combined along the feature axis, f(·) is the activation function, K is the convolution kernel parameter, b is the bias, and s is the distance that the convolution moves along the time axis at a specific step size. is the temporal convolution matrix.

5. The method for controlling the safe and stable voltage of a power grid based on TCRN and GOA according to claim 4, characterized in that: The construction of the voltage stability prediction model also includes using a short-circuit method to perform feature extraction, which is expressed as: H(x)=x+F(x) Among them, F(x) is the output feature of input x after weighting, and H(x) is the output feature using short-circuit wires; The activation unit is modified, expressed as: ReLU(x in )=max(0,x in ) Among them, ReLU is the rectified linear activation function, x in Input for the rectified linear activation unit; The batch normalization layer is used to optimize the network training efficiency, which can be expressed as: Among them, x is the network adaptive change input, is the normalized network adaptive change output, m is the number of samples, γ and β are the scaling coefficient and translation coefficient of the network respectively, and ε is a positive number to prevent the denominator from being zero.

6. The method for controlling the safe and stable voltage of a power grid based on TCRN and GOA according to claim 5, characterized in that: The construction of the voltage stability control model includes constructing a voltage stability control model based on the Pelican optimization algorithm, taking the machine cutting and load cutting capacity as the optimization target, and constructing a control model with the minimum machine cutting and load cutting capacity, which is expressed as: Among them, ΔP G，i is the generator removal capacity of the i-th node, ΔP L，j u j is the load shedding capacity of the jth node, N G 、N L are the number of generator and load shedding nodes, u i 、u j are generator and load shedding factors respectively; Load shedding factor u j is a continuous variable between 0 and 1, and the generator removal coefficient u is calculated. i , expressed as: Setting voltage stability evaluation index ξ m Above the threshold ξ, the constraint formula is expressed as: Among them, ξ k，0 is the voltage stability evaluation index of the kth node when no stability control measures are implemented after a large interference fault, λ k，i is the voltage stability control sensitivity coefficient, λ k，j Adjust sensitivity factor for load; Calculate the voltage stability control sensitivity coefficient, expressed as: Among them, ξ k is the adjusted voltage stability evaluation index value, ΔP i is the power removal amount of the i-th generator; Define the constraints of power balance, expressed as: Among them, S b,min、 S b,max They are respectively the minimum and maximum values ​​of the power unbalance capacity when the generator frequency regulation factor is introduced.

7. The method for controlling the safe and stable voltage of a power grid based on TCRN and GOA according to claim 6, characterized in that: The outputting of the optimal machine cutting and load shedding strategy includes setting the machine cutting and load shedding capacity to individuals of the Pelican optimization algorithm, initializing the Pelican optimization algorithm parameters including the Pelican position, speed, number of iterations, and population size, and adjusting the machine cutting and load shedding capacity according to the Pelican position; According to the system voltage node bus sequence, the voltage stability prediction based on the time series convolution residual network is obtained. If the voltage is stable, the voltage stability control model objective function is calculated, and the optimal position of the pelican is determined according to the objective function, and the pelican position is updated; If the voltage is unstable, the pelican position is not updated; When all pelicans’ positions are updated, if convergence is achieved or the maximum number of iterations is reached, the final machine shedding capacity is output; If convergence is not met or the maximum number of iterations is not reached, the load shedding capacity of the cutting machine is re-adjusted according to the position of the pelican.

8. A system using the TCRN and GOA-based grid voltage safety and stability control method as claimed in any one of claims 1 to 7, characterized in that: It includes voltage stability assessment module, voltage stability prediction module and voltage stability control optimization module; The voltage stability assessment module is used to refine the grid stability impact by assigning voltage reduction interval weights and establish a voltage stability assessment quantitative index; The voltage stability prediction module is used to construct a voltage stability prediction model based on the voltage stability evaluation quantitative index to predict the relationship between voltage state and stability; The voltage stability control optimization module is used to construct a voltage stability control model through prediction results and output optimal generator and load shedding strategies.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the grid voltage security and stability control method based on TCRN and GOA described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the grid voltage security and stability control method based on TCRN and GOA described in any one of claims 1 to 7 are implemented.

Images