A power system voltage stability evaluation method based on improved cascade forest

Abstract

A power system voltage stability evaluation method based on improved cascade forest, step 1: build fast voltage stability index FVSI to describe the power system voltage stability state, obtain the power system voltage stability margin; step 2: the mapping relationship between input variables and output variables is represented by using an improved cascade forest Improved Cascade Forest model; step 3: the improved cascade forest model is optimized by using an improved zebra optimization algorithm Improved Zebra Optimization Algorithm; step 4: the data set of the IZO A improved ICF model is constructed by using an offline power flow method, the data is pretreated, and an efficient data set is generated; step 5: the efficient data set is divided into a training set and a test set, the IZO A optimized ICF model is trained by using the training set, and the model is tested by using the test set. The present application is suitable for new power systems with new energy access, and has important significance for guaranteeing the safe and stable operation of the power system.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of power system voltage stability and new energy voltage, and particularly relates to a power system voltage stability evaluation method based on an improved cascade forest. TECHNICAL BACKGROUND

[0002] In recent years, with the increasingly serious energy problem, green low-carbon energy represented by wind energy and solar energy has been vigorously developed. The access of new energy such as wind energy and solar energy poses higher challenges to power system voltage stability. Voltage stability problem is one of the important considerations in power system planning, operation and design. Continuous disturbance of the power system will cause voltage instability, and sometimes even lead to failure of the entire system. The main reason for voltage instability of the power system is insufficient supply of reactive power or excessive occupation of reactive power by the system itself. In order to ensure the safe operation of the system, it is of great significance to evaluate the voltage stability of the system.

[0003] The document with application publication number CN108988320A discloses a method for analyzing the influence of dynamic element response characteristics of a power system on voltage stability. From the perspective of generalized branch potential energy, a new analysis method is proposed to study the influence of generators, high-voltage direct current transmission systems and induction motors on voltage stability. According to the change rule of transient potential energy distribution in the network after fault and the information of stability margin, the reactive power recovery characteristics of each dynamic element after fault are taken as the basic variables, an evaluation index is established, and the mechanism of causing transient voltage instability of the power grid is analyzed; the document with application publication number CN110233476A discloses a voltage stability evaluation method in black start process and related system. The specific steps are as follows: analyzing several factors affecting voltage stability in the black start process and sorting them according to the influence degree, selecting necessary elements as feature inputs of the support vector machine model, and selecting several typical values for each feature input to form a sample set. The sample set is used to train the support vector machine model to obtain the mapping relationship between the feature input and the output, and a voltage stability evaluation model is constructed. The actual parameters in the black start process and the voltage stability evaluation model are used to evaluate the voltage stability; the document with application publication number CN114626757A discloses a method and system for judging voltage stability of a receiving end system. The specific steps are as follows: obtaining bus transient voltage stability data, and using difference equations to segment the voltage curve. According to the existing transient voltage practical criterion, different weights are given according to the different voltage drop degrees. The method of weighted integral is used to quantify the drop accumulation. An evaluation index is constructed to form a complete discrimination method.

[0004] The purpose of the present application is to solve the technical problems of low accuracy, complex calculation and long time consumption in voltage stability evaluation of power system in the technical literature, to introduce artificial intelligence algorithm into the field of voltage stability evaluation of power system, and to optimize the improved cascade forest model by using the improved zebra optimization algorithm, and to propose the power system voltage stability evaluation method based on improved cascade forest. SUMMARY

[0005] The purpose of the present application is to solve the technical problems of low accuracy, complex calculation and long time consumption in voltage stability evaluation of power system in the technical literature, to introduce artificial intelligence algorithm into the field of voltage stability evaluation of power system, and to optimize the improved cascade forest model by using the improved zebra optimization algorithm, and to propose the power system voltage stability evaluation method based on improved cascade forest.

[0006] The technical scheme adopted by the present application is:

[0007] A voltage stability evaluation method based on improved cascade forest, comprising the following steps:

[0008] Step 1: Construct a fast voltage stability index (FVSI) to describe the voltage stability state of the power system, and obtain the voltage stability margin of the power system;

[0009] Step 2: Use an improved cascade forest (ICF) model to represent the mapping relationship between input variables and FVSI;

[0010] Step 3: Use an improved zebra optimization algorithm (IZOA) to optimize the ICF model;

[0011] Step 4: Use an offline power flow method to build a data set for the IZA improved ICF model, preprocess the data, and generate an efficient data set;

[0012] Step 5: Divide the efficient data set into training set and test set, use the training set to train the ICF model optimized by IZA, use the test set to test the model, and judge the voltage stability of the power system.

[0013] In step 1: Construct FVSI to describe the voltage state of the power system, and obtain the voltage stability margin of the power system, which specifically includes the following steps:

[0014] Step 1-1: Construct FVSI to describe the voltage state of the power system;

[0015] The voltage stability index can be used to study the voltage stability of the power system. The FVSI is used to detect the key bus in the power system, and the FVSI calculation formula is as follows:

[0016]

[0017] In the formula: Z sr is the line impedance from the sending end to the receiving end; Q r is the receiving end reactive power; V s is the sending end voltage; X sr is the line reactance from the sending end to the receiving end.

[0018] Step 1-2: Obtain the voltage stability margin of the power system based on FVSI.

[0019] In order to make the power system operate stably, the FVSI value of the power system needs to be kept in a range much smaller than 1.

[0020] When the FVSI value is close to 1, it indicates that the line is close to instability, and in severe cases, it may cause the voltage collapse of the power system.

[0021] In step 2: the ICF model is used to represent the mapping relationship between the input variables and the voltage stability margin, which includes the following steps:

[0022] Step 2-1: Based on the traditional Cascade Forest (CF) model, the ICF model is obtained by improvement;

[0023] The CF model structure is improved by increasing the types of base learners. The CatBoost, XGBoost, Random Forest and LightGBM learners are integrated to obtain the ICF model.

[0024] Step 2-2: Based on the ICF model, the mapping relationship between the input variables and the FVSI is represented.

[0025] The bus voltage amplitude, generator active and reactive power and other input variables are input into the ICF model. Each layer uses CatBoost, XGBoost, Random Forest and LightGBM learners to train the model respectively, obtains the prediction probability, and adds these probabilities as new features to the input variables of the next layer. Each layer of the model performs feature transformation and fusion, and finally, through the multi-level cascade model, the accurate FVSI can be continuously optimized.

[0026] In step 3, the IZOA is used to optimize the ICF model, which includes the following steps:

[0027] Step 3-1: Set algorithm parameters, including the number of zebra populations, iteration convergence conditions, minimum and maximum values of parameters to be optimized in the ICF model, etc.

[0028] Step 3-2: Introduce Elite Opposition-Based Learning (EOBL) to initialize the population: First, randomly generate a zebra population, then compare the fitness function value of the existing zebra with that of the elite opposition zebra, and select the one with the lower fitness function value as the next generation zebra.

[0029] Each zebra i is defined as a position vector Z i = (z i,1 ,···,z i,j ,···,z i,m ), m is the value of the hyperparameter to be optimized in the ICF model in step 2, and the fitness function formula is as follows:

[0030] F(Z i ) = min(1 - Acc CF ) (9)

[0031] In the formula, Acc CF represents the accuracy of the ICF model.

[0032] The zebra with the lowest fitness function value in the population is defined as the pioneer zebra, which guides other zebras to its position during the optimization process.

[0033] EOBL is an optimization strategy that uses elite individuals to construct an opposition population to increase population diversity. For any zebra Z i , its opposition zebra is represented as

[0034]

[0035] In the formula, r is a random number between 0 and 1; LB is the lower bound of the elite zebra; and UB is the upper bound of the elite zebra.

[0036] Step 3-3: Foraging stage position update: calculate the new position of the zebra using formula (3), and update the position using formula (10);

[0037] The first stage updates the population individuals by simulating the behavior of zebras during foraging. The optimal individual in the population is called the pioneer zebra, which guides the remaining individuals to move towards its position within the search area. The zebra position update formula is as follows:

[0038]

[0039]

[0040] wherein: is the updated position of zebra i in the jth dimension after foraging behavior; z i,j is the position of zebra i in the jth dimension; MZ j is the optimal position in the jth dimension; r is a random number between 0 and 1; l e {1, 2}, the greater the value of l, the faster the movement of the population changes; is the new position of zebra i after foraging behavior; Z i is the actual position of zebra i; F i P1 is the fitness value of the target position after foraging behavior; F i is the fitness value before movement.

[0041] Step 3-4: Position update in defense stage: according to the value of random number p s , the new position of the zebra is calculated using formula (3), and the position is updated using formula (11).

[0042] In the second stage, the population position is updated using the defense behavior of zebras facing predators. Zebras have different defense behaviors when facing different predators: when attacked by lions, zebras will choose to escape as shown in formula (4) S1; when facing other predators, zebras will choose to attack as shown in formula (4) S2.

[0043]

[0044]

[0045] wherein: is the updated position of zebra i in the jth dimension after defense behavior; is the new position of zebra i after defense behavior; F i P2 is the fitness value of the target position after defense behavior; t and T are the current iteration number and the total number of iterations, respectively; k is a constant of 0.01; P s is a random number between 0 and 1; NZ j is the optimal position in the jth dimension.

[0046] Step 3-5: Introduce firefly disturbance strategy (FDS) to update zebra position again: calculate the new position of zebras using formula (5), then compare the fitness function value of the existing zebra with the fitness function value of the pioneer zebra, and select the one with lower fitness function value as the next generation zebra;

[0047]

[0048] ​​

[0049] where: β0 and θ are constants; S is the spatial distance between the zebra and the pioneer zebra; δ is a step factor belonging to the interval [0, 1].

[0050] Step 3-6: Verify whether the iteration limit is met: if the maximum number of iterations is not reached, return to step 3-2 to continue iteration. If it is reached, end the optimization process.

[0051] In step 4: use the off-line power flow method to build the data set of the IZO A improved ICF model, preprocess the data, and generate an efficient data set, which includes the following steps:

[0052] Step 4-1: Normalize the data based on the MAXABS method;

[0053] The MAXABS normalization method rescales the data x to the range [-1, 1], and the calculation formula is as follows:

[0054]

[0055] where: x is the data; |x max is the maximum absolute value of the data.

[0056] Step 4-2: Use the Fisher Score algorithm to filter the original data to obtain an efficient data set.

[0057] Step 4-2-1: Mark the state of the data normalized in step 4-1, and divide the data into a training data set and a test data set according to a certain proportion;

[0058] Step 4-2-2: Use the Fisher Score algorithm to calculate the importance of the electrical features in the original data;

[0059] Assume that the number of stable and unstable samples in all electrical feature samples is K s and K m , respectively, and the formula for calculating the score F a of the a-th feature of the sample using the Fisher Score algorithm is as follows:

[0060]

[0061] where: m n , and are the mean values of the a-th feature of the sample in the entire data set, in the stable sample data set, and in the unstable sample set, respectively; is the feature value of the a-th feature of the i-th stable sample data; is the eigenvalue of the a-dimensional feature of the i-th unstable sample data.

[0062] Step 4-2-3: arrange all features in descending order according to Fisher Score values;

[0063] Step 4-2-4: add the features to the training data set in sequence to train the ICF model, and obtain the feature dimension when the model accuracy is the maximum.

[0064] In step 5: divide the high-efficiency data set into a training set and a test set, train the ICF model optimized by IZO A using the training set, test the model using the test set, and judge the voltage stability of the power system.

[0065] Compared with the prior art, the beneficial effects of the present application are that:

[0066] 1. To solve the influence of new energy on the voltage stability of the power system, the present application adopts the ICF model and proposes a simple, fast and computationally feasible method for evaluating the voltage stability of the power system.

[0067] 2. The present application uses Fisher Score algorithm to filter the original data to obtain a high-efficiency data set. Fisher Score algorithm is a statistical method for feature selection, which is simple to understand, high in computational efficiency and wide in adaptability.

[0068] 3. The present application improves the zebra optimization algorithm, the EOBL initialization population can improve the global search ability, convergence accuracy and speed of the optimization algorithm, while avoiding falling into local optimum, accelerating the search ability of the algorithm, and using FDS to update the zebra position, which improves the search ability of the algorithm and avoids premature convergence. BRIEF DESCRIPTION OF DRAWINGS

[0069] The present application will be further described below in combination with the drawings and examples:

[0070] Figure 1 is the flow chart of the method of the present application;

[0071] Figure 2 is the improved cascade forest model structure of the present application;

[0072] Figure 3 is the improved zebra optimization algorithm flow chart of the present application;

[0073] Figure 4 is the schematic diagram of the IEEE 14-node system of the present application. DETAILED DESCRIPTION

[0074] The application provides a power system voltage stability evaluation method based on an improved cascade forest, and the method comprises the following steps, as shown in the specification: Figure 1

[0075] Step 1: constructing a fast voltage stability index (FVSI) to describe the voltage stability state of a power system, so as to obtain a power system voltage stability margin;

[0076] Step 2: using an improved cascade forest (ICF) model to represent the mapping relationship between input variables and the FVSI;

[0077] Step 3: using an improved zebra optimization algorithm (IZOA) to optimize the ICF model;

[0078] Step 4: using an offline power flow method to construct a data set of the IZOA improved ICF model, pre-processing the data, and generating an efficient data set;

[0079] Step 5: dividing the efficient data set into a training set and a test set, using the training set to train the ICF model optimized by the IZO A, using the test set to test the model, and judging the voltage stability of the power system.

[0080] In step 1: constructing the FVSI to describe the voltage state of the power system, so as to obtain the power system voltage stability margin, which specifically comprises the following steps:

[0081] Step 1-1: constructing the FVSI to describe the voltage state of the power system;

[0082] The voltage stability index can be used to study the voltage stability of the power system. The FVSI is used to detect the key bus in the power system, so as to obtain the voltage stability margin of the power system, and the FVSI calculation formula is as follows:

[0083]

[0084] In the formula, Z sr is the line impedance from the sending end to the receiving end; Q r is the receiving end reactive power; V s is the sending end voltage; X sr is the line reactance from the sending end to the receiving end.

[0085] Step 1-2: obtaining the power system voltage stability margin based on the FVSI.

[0086] ​In order to make the power system run stably, the FVSI value of the power system needs to be kept in a range much less than 1. When the FVSI value is close to 1, it indicates that the line is close to instability, and in severe cases, it may cause the voltage collapse of the power system.

[0087] In step 2: the mapping relationship between the input variables and the voltage stability margin is represented by the ICF model, which specifically includes the following steps:

[0088] Step 2-1: based on the traditional Cascade Forest (CF) model, an ICF model is obtained by improvement;

[0089] The structure of the CF model is improved, and the types of base learners are increased. The CatBoost, XGBoost, Random Forest, and LightGBM learners are integrated to obtain the ICF model.

[0090] Step 2-2: the mapping relationship between the input variables and the FVSI is represented based on the ICF model.

[0091] The bus voltage amplitude, generator active and reactive power, and other input variables are input into the ICF model. Each layer uses CatBoost, XGBoost, Random Forest, and LightGBM learners to train the model respectively, obtains the prediction probability, and adds these probabilities as new features to the input variables of the next layer. Each layer of the model performs feature conversion and fusion, and finally, through the multi-level cascade model, the accurate FVSI can be continuously optimized.

[0092] In step 3, the ICF model is optimized by IZO A, the training process is completed, and the optimal weight and threshold are found, as shown in Figure 3 , which specifically includes the following steps:

[0093] Step 3-1: set the algorithm parameters, including the number of zebra populations, the iteration convergence condition, the minimum and maximum values of the parameters to be optimized in the ICF model, etc.;

[0094] Step 3-2: introduce Elite Opposition-Based Learning (EOBL) to initialize the population: first, randomly generate a zebra population, then compare the fitness function values of the existing zebra and the elite opposition zebra, and select the one with lower fitness function value as the next generation zebra;

[0095] Each zebra i is defined as a position vector Z i =(z i,1 ,···,z i,j ,···,z i,m), m is the value of the hyper-parameter to be optimized in the improved CF model in step 2, and the fitness function is as follows:

[0096] F(Z i ) = min (1 - Acc CF ) (9)

[0097] Acc CF represents the accuracy of the improved CF model.

[0098] The zebra with the lowest fitness function value in the population is defined as the pioneer zebra, which guides other zebras to its position during the optimization process.

[0099] EOBL is an optimization strategy that uses elite individuals to construct a reverse population to increase the diversity of the population. For any zebra Z i , its reverse zebra is represented as

[0100]

[0101] where r is a random number between 0 and 1; LB is the lower bound of the elite zebra; and UB is the upper bound of the elite zebra.

[0102] Step 3-3: Position update in the foraging stage: calculate the new position of the zebra using formula (3), and update the position using formula (10);

[0103] The first stage updates the population individuals by simulating the behavior of zebras during foraging. The best individual in the population is called the pioneer zebra, which guides the remaining individuals to its position within the search area. The zebra position update formula is as follows:

[0104]

[0105]

[0106] where: is the updated position of zebra i in the jth dimension after foraging behavior; z i,j is the position of zebra i in the jth dimension; MZ j is the optimal position in the jth dimension; r is a random number between 0 and 1; l ∈ {1, 2}, the larger the value of l, the faster the movement of the population; is the new position of zebra i in the foraging behavior; Z i is the actual position of zebra i; F i ′P1 is the fitness value at the target position of foraging behavior; F i is the fitness value before movement.

[0107] Step 3-4: Defense stage position update: according to the value of random number p s , the new position of zebra is calculated using formula (4), and the position is updated using formula (11).

[0108] In the second stage, the population position is updated using the defense behavior of zebras against predators. Zebras have different defense behaviors against different predators: when attacked by lions, zebras will choose to escape as shown in formula (4) S1; when facing other predators, zebras will choose to attack as shown in formula (4) S2.

[0109]

[0110]

[0111] In the formula: is the updated position of zebra i in the jth dimension after defense behavior; is the new position of the ith zebra after defense behavior; F i ′P2 is the fitness value on the target position after defense behavior; t and T are the current iteration number and total number of iterations, respectively; k is a constant 0.01; P s is a random number between 0 and 1; NZ j is the optimal position in the jth dimension.

[0112] Step 3-5: Introduce firefly disturbance strategy (FDS) to update zebra position again: calculate the new position of zebras using formula (5), then compare the fitness function value of the existing zebra with the fitness function value of the pioneer zebra, and select the one with lower fitness function value as the next generation zebra;

[0113]

[0114]

[0115] In the formula: β0 and θ are constants; S is the spatial distance between zebra and pioneer zebra; δ is a step factor belonging to the interval [0, 1].

[0116] Step 3-6: Verify whether the iteration limit is met: if the maximum number of iterations is not reached, return to step 3-2 to continue iteration. If it is reached, end the optimization process.

[0117] In step 4: offline power flow method is used to build the data set of IZO A improved ICF model, and the data is preprocessed to generate efficient data set, which includes the following steps:

[0118] Step 4-1: Normalize the data based on the MAXABS method;

[0119] MAXABS normalization method rescales the data x to the range [-1, 1] and the calculation formula is as follows:

[0120]

[0121] In the formula: x is the data; |x max is the maximum value of the absolute value of the data.

[0122] Step 4-2: The Fisher Score algorithm is used to filter the original data to obtain a high-efficiency data set.

[0123] Step 4-2-1: State marking is performed on the normalized data of step 4-1, and the data is divided into a training data set and a test data set according to a certain proportion;

[0124] Step 4-2-2: The Fisher Score algorithm is used to calculate the importance of the electrical characteristics in the original data;

[0125] Suppose the number of stable and unstable samples in all electrical characteristic samples is K s and K m , respectively, and the Fisher Score algorithm is used to calculate the score F a of the a-th feature of the sample. The formula is as follows:

[0126]

[0127] In the formula: m n , and are the mean values of the a-th feature of the sample in the entire data set, in the stable sample data set, and in the unstable sample set, respectively; is the feature value of the a-th feature of the i-th stable sample data; is the feature value of the a-th feature of the i-th unstable sample data.

[0128] Step 4-2-3: All features are arranged in descending order according to the Fisher Score value;

[0129] Step 4-2-4: The features are sequentially added to the training data set to train the WNN model, and the feature dimension when the model accuracy is maximum is obtained by comparison.

[0130] In step 5: The high-efficiency data set is divided into a training set and a test set, the IZO A optimized ICF model is trained using the training set, the model is tested using the test set, and the voltage stability of the power system is determined.

[0131] Embodiment:

[0132] To verify the effectiveness of the present application, tests are carried out on an IEEE 14-node system, and a schematic diagram of the IEEE 14-node system is as shown in Figure 4 In the IEEE 14-bus system, the synchronous generators in bus 6 and bus 2 are replaced by a wind farm (WF) based on a doubly-fed induction generator, and the accuracy of the improved cascade forest model based on the improved zebra optimization algorithm in voltage stability evaluation is verified through simulation.

[0133] After the ICF model is trained by using the training set, the ICF model is tested by using a test set under various unknown load conditions. The experimental results of the continuation power flow (CPF) method based on the FVSI value are verified. Under different load conditions, the FVSI values of the IEEE 14-node system measured by the two methods are obtained, and the minimum bus and the maximum bus are obtained by comparing the FVSI values of the buses. Under the reference load condition, the minimum bus and the maximum bus of the FVSI are as shown in Table 1; under the maximum load condition, the minimum bus and the maximum bus of the FVSI are as shown in Table 2.

[0134] Table 1

[0135]

[0136] Table 2

[0137]

[0138] As can be seen from the above table, compared with the CPF method, the trained ICF gives accurate FVSI values, and the FVSI value of bus 14 is closer to 1 than that of other buses, which is more close to voltage instability. The FVSI value of bus 4 is low, indicating that bus 4 has stronger voltage stability. Table 1 and Table 2 respectively give the FVSI values under the reference load condition and the maximum load condition. In the two scenarios, bus 14 is determined as a voltage weak bus, and a compensation device is placed to improve the voltage stability of the entire network. The method proposed in the present application is compared with the CPF method, and the improved cascade forest model is used to evaluate the voltage stability of the power system, which is not only fast but also approximately accurate.

[0139] The power system voltage stability evaluation method based on the improved cascade forest proposed in the present application has important significance for ensuring the safe and stable operation of the power system.

Claims

1. A power system voltage stability assessment method based on improved cascade forest, characterized in that, The method comprises the following steps: Step 1: Constructing a fast voltage stability index Describing the voltage stability state of the power system, obtaining the voltage stability margin of the power system; Step 2: Represent the mapping relationship between the input variables and the fast voltage stability index using the improved cascade forest ICF model ; Step 3: the improved zebra optimization algorithm IZO A is used to optimize the improved cascade forest ICF model; Step 4: the off-line power flow method is used to build a data set of the algorithm-optimized improved cascade forest ICF model, the data is preprocessed, and an efficient data set is generated; Step 5: the efficient data set is divided into a training set and a test set, the training set is used to train the model, and the test set is used to test the model; In step 3, the improved zebra optimization algorithm IZO A is used to optimize the improved cascade forest ICF model, which specifically comprises the following steps: Step 3-1: set the algorithm parameters, including the number of zebra populations, the iteration convergence condition, and the minimum and maximum values of the parameters to be optimized in the improved cascade forest ICF model; Step 3-2: introduce the elite opposition learning EOBL to initialize the population: first, randomly generate the zebra population, then compare the fitness function value of the existing zebra with the fitness function value of the elite opposition zebra, and select the one with the lower fitness function value as the next generation zebra; Step 3-3: update the position in the foraging stage: calculate the new position of the zebra and update the position of the zebra; Step 3-4: Defense Phase Location Update: Based on the value of the random number, the new location of the zebra is calculated. Step 3-4: Defense Phase Location Update: Based on the value of the random number, the new location of the zebra is calculated. Step 3-5: introduce the firefly disturbance strategy FDS to update the zebra position again: calculate the new position of the zebra, then compare the fitness function value of the existing zebra with the fitness function value of the pioneer zebra, and select the one with the lower fitness function value as the next generation zebra; Step 3-6: verify whether the iteration limit is met: if the maximum iteration number is not reached, return to step 3-2 for iteration; if the maximum iteration number is reached, end the optimization process.

2. The method of claim 1, wherein: In step 1, a fast voltage stability index is constructed Describing the voltage state of the power system, obtaining the power system voltage stability margin, specifically comprising the following steps: Step 1-1: Constructing a fast voltage stability index Describing the voltage state of a power system; The voltage stability index can be used to study the voltage stability of the power system, and the fast voltage stability index The key bus in the power system is detected, and the calculation formula is as follows: （1）； wherein: Z is the line impedance from the sending end to the receiving end; Qr is the reactive power at the receiving end; Vs is the sending end voltage; X is the line reactance from the sending end to the receiving end; Steps 1-2: Based on fast voltage stability index Obtaining the voltage stability margin of the power system; Fast voltage stability index The smaller the value, the more stable the line is. When the value is close to 1, it indicates that the line is close to instability, and in severe cases, it may lead to voltage collapse of the power system.

3. The method of claim 1, wherein: In step 2, an improved cascaded forest ICF model is used to represent the input variables and the fast voltage stability index. The mapping relationship between them specifically includes the following steps: Step 2-1: improve the traditional cascade forest CF model to obtain the improved cascade forest ICF model; The structure of the cascade forest CF model is improved, and the types of base learners are increased; CatBoost, XGBoost, random forest and LightGBM learners are integrated to obtain the improved cascade forest ICF model; Step 2-2: Representing the mapping relationship between the input variables and the fast voltage stability index based on the improved cascade forest ICF model ; The bus voltage amplitude, generator active and reactive power are input variables into the improved cascaded forest ICF model. Each layer uses CatBoost, XGBoost, Random Forest, and LightGBM learners to train the model and obtain prediction probabilities. These probabilities are then added as new features to the input variables of the next layer. Feature transformation and fusion are performed at each layer. Finally, through multi-level cascaded models, accurate and fast voltage stability indicators can be obtained through continuous optimization. .

4. The method of claim 1, wherein: In step 3-2, the elite opposition-based learning (EOBL) initializes the population, and the reverse population is constructed by using the elite individuals, specifically: for any zebra , the reverse zebra is expressed as , and the formula is as follows: （2）； wherein: is a random number between 0 and 1 ; is a lower bound for elite zebras; is an upper bound for elite zebras; In step 3-3, the new position of the zebra is calculated using formula (3), and the position of the zebra is updated; The zebra position updating formula after the foraging stage is as follows: （3）； wherein: is the foraging behavior posterior zebra In the first dimension update position; is the foraging behavior posterior zebra In the first dimension position; is the first dimension optimal position; is a random number between 0 and 1; , The greater the value, the faster the population moves. In step 3-4, when attacked by a lion, the zebra will choose an escape strategy as shown in equation (4) ; when facing other predators, the zebra will choose an attack strategy as shown in equation (4) ; （4）； wherein: to defend against the behavior of the zebra In the first dimensional update position; is a random number between 0 and 1; and are the current iteration number and the total number of iterations, respectively; is a constant 0.01; is a random number between [0, 1]; is the first dimensional optimal position; In step 3-5, the new position of the zebra is calculated using formula (5), then the fitness function value of the existing zebra is compared with the fitness function value of the pioneer zebra, and the one with the lower fitness function value is selected as the next generation zebra; （5）； （6）； wherein: and are constants; is the spatial distance between the zebra and the pioneer zebra; is a step factor belonging to the interval .

5. The method of claim 1, wherein: In step 4, the off-line power flow method is used to build a data set of the algorithm-optimized improved cascade forest ICF model, the data is preprocessed, and an efficient data set is generated, which specifically comprises the following steps: Step 4-1: normalize the data based on the MAXABS method; The MAXABS normalization method rescales the data to the range [0, 1] with the following formula: MAXABS = 1 - MINABS （7）； In the formula: is data; is the maximum absolute value of the data Step 4-2: filter the original data using the Fisher Score algorithm to obtain an efficient data set; Step 4-2-1: mark the state of the data normalized in step 4-1, and divide the data into a training data set and a test data set according to a certain proportion; Step 4-2-2: calculate the importance of the electrical characteristics in the original data using the Fisher Score algorithm; Assume that the number of stable and unstable samples in all electrical feature samples is and , the formula for calculating the score of the th feature of the sample is as follows: ​ （8）； In the formula: , and The samples are respectively The dimensional feature is the mean of the entire dataset, the stable sample dataset, and the unstable sample set; For the first The first stable sample data Eigenvalues ​​of a dimensional feature; For the first The first unstable sample data Eigenvalues ​​of a dimensional feature; Step 4-2-3: sort all the characteristics in descending order according to the Fisher Score values; Step 4-2-4: The features are added to the training data set in sequence to train the improved cascade forest ICF model, and the feature dimension when the model accuracy is the largest is obtained by comparison.

Images