Method for determining sensitive points of power angle stability control of electric power system and related equipment

Abstract

The invention relates to the technical field of power system control, and discloses a method for determining power angle stability control sensitive points of a power system and related equipment, and the method comprises the steps: obtaining the power angle stability margin improvement amount under different control measures based on a generated system power angle stability margin evaluation model under the occurrence of a power angle instability phenomenon; calculating sensitivity screening effective control points through the power angle stability margin lifting amount, determining control sensitive points according to a sensitivity sorting result, and completing the determination work of the power angle stability control sensitive points of the electric power system. According to the method, the power angle stability margin lifting amount under different control measures is calculated, the effective control points are screened through the sensitivity, and finally the control sensitive points are determined. According to the method, the control point which has the greatest influence on the stability of the power system can be accurately identified, so that control measures are specifically applied, and the control efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power system control, and specifically to a method for determining sensitive points of power angle stability control in a power system and related equipment. Background Art

[0002] With the advancement of the goal, the process of green transformation of China's energy has accelerated, and the power system shows the "dual-high" characteristics of a high proportion of renewable new energy and a high proportion of power electronics. However, with a large number of new energy power stations incorporated into the backbone grid, the safe and stable operation of the power grid faces huge challenges.

[0003] Power system transient stability refers to the ability of a power system to maintain synchronous operation and transition to a new or restore to the original steady-state operation mode after being subjected to a large disturbance. The large-scale access of new energy and the deepening of the AC-DC hybrid connection have greatly increased the complexity of the grid structure and operation, resulting in fundamental changes in the system dynamics characteristics and stability mechanisms, leading to problems such as reduced system inertia and decreased anti-interference ability in the power system, seriously threatening the transient power angle stability of the sending-end system.

[0004] Currently, in order to reduce the risk of power system instability under large disturbances and ensure the global stable operation of the system, transient stability emergency control measures such as generator tripping, load shedding, DC transmission emergency power modulation, and local islanding are often adopted. Among them, generator tripping is the simplest and most mature control measure with the widest application range.

[0005] Traditional generator tripping scheme formulation can be assisted by time-domain simulation methods, control optimization methods, and sensitivity-based methods. In the methods based on time-domain simulation and engineering experience, the time-domain changes or stable performance index information of key electrical quantities of each generator are obtained as the basis for the generator tripping strategy. However, the trial-and-error cost of formulating the scheme based on engineering experience is too high, and the repeated time-domain simulation calculations are time-consuming and the calculation accuracy cannot be guaranteed. In the optimization-based methods, in order to obtain the optimal control quantity, it is necessary to solve non-linear optimization equations, which are time-consuming and not easy to converge. For the sensitivity-based methods, by evaluating the improvement effect of each control point on the stability margin, the generator tripping point with high control sensitivity is selected as the action point. Although the application of sensitivity can reduce the computational complexity, in different operating scenarios, there may be problems such as unclear mapping relationship between the stability margin and sensitivity, and insufficient calculation accuracy of the control measure sensitivity. Therefore, it is urgent to accurately evaluate the enhancement effect of each controllable generator tripping resource on the power angle stability margin in a high-proportion renewable new energy power system under different operating modes, so as to provide a basis for formulating optimized control strategies. Summary of the Invention

[0006] To overcome the defects existing in the above-mentioned prior art, the purpose of the present invention is to provide a method for determining sensitive points of power angle stability control in a power system and related equipment, so as to solve the technical problem of how to improve the accuracy of evaluating the enhancement effect of each controllable generator tripping resource on the power angle stability margin in a proportional renewable new energy power system under different operating modes in the prior art.

[0007] The present invention is realized through the following technical solutions: In the first aspect, the present invention provides a method for determining sensitive points of power angle stability control in a power system, including: Performing time-domain simulation based on the operating mode and fault form of the power system to generate a basic scenario; Collecting the simulation data output by the basic scenario, calculating the power angle stability margin, judging the stability level of the system through the power angle stability margin. When the system is unstable, selecting potential generator tripping control points, applying control measures to form power angle stability samples, and combining the power angle stability samples with the basic scenario to generate a power angle stability sample set; Using the key electrical characteristic time series curve and control variables as input features, and the corresponding power angle stability margin as the output label, improving and training the CNN model to determine the optimal structure and parameters of the model, and generating a system power angle stability margin evaluation model; Based on the system power angle stability margin evaluation model, obtaining the power angle stability margin improvement amount under different control measures in the case of power angle instability, calculating the sensitivity through the power angle stability margin improvement amount to screen effective control points, and determining the control sensitive points according to the sensitivity sorting result, completing the determination work of the power angle stability control sensitive points in the power system. Preferably, the simulation data output by the basic scenario includes the power angle time series curve and key electrical parameter time series information under different operating scenarios.

[0008] Preferably, in the step of judging the stability level of the system through the power angle stability margin, the extended equal area criterion is used for stability analysis, including: Dividing the multi-machine system into a critical machine group and a remaining machine group, and then equivalent the system to a two-machine system, i.e., a single-machine infinite bus system, according to the inertia centers of the critical machine group and the remaining machine group; Based on the power angle time series curve before and after fault excision, calculating the accelerating area of the power angle time series curve S A and the decelerating area S B , then through the accelerating area S A and the decelerating area S B Calculating the system power angle stability margin, and the formula is as follows:

[0009] Among them, , representing the deceleration area S B is greater than or equal to the acceleration area S A , the system is stable; , representing the deceleration area S B is less than the acceleration area S A , the system is unstable; The larger it is, the greater the difference between the deceleration area and the acceleration area of the system, and the greater the system stability margin; The smaller it is, the smaller the difference between the deceleration area and the acceleration area of the system, and the smaller the system stability margin.

[0010] Preferably, the CNN model includes an input layer, a convolutional layer, a downsampling layer, a fully connected layer, and an output layer; The process of improving the CNN model is as follows: Adopt multiple convolutional layers to extract multi-dimensional features, and reduce the size of the convolutional kernel to reduce the number of calculation parameters; Use the unsaturated non-linear function Leaky ReLU as the activation function to fit various features; Regularize the loss function to correct the over-extraction of feature variables and reduce the model complexity.

[0011] Preferably, for the generated system power angle stability margin evaluation model, a predictive evaluation is carried out, and the root mean square error is used to calculate the sensitivity numerical prediction error. The calculation formula is as follows:

[0012] Among them, N is the number of samples, y' ( i ) and y ( i ) are the predicted value and the true value of the i th sample respectively.

[0013] Preferably, in the step of screening effective control points by calculating the sensitivity of the power angle stability margin improvement amount, the process of determining the sensitivity is as follows: Establish the correlation relationship between the control measure and the change amount of the transient stability margin, and the expression is as follows: ; In the formula: represents the transient stability margin of the system after taking emergency control measures, represents the transient stability margin of the system without taking emergency control measures, represents thek The control adjustment amount of a control point; Determine the sensitivity based on the correlation between the control measure and the change in transient stability margin. The calculation formula is as follows: ; Wherein, represents the control adjustment amount of the k th control point; represents the correlation between the control measure and the change in transient stability margin; S k characterizes the influence of the control point k on the power angle stability margin of the system, representing the power angle stability margin changed by the unit control amount of the control point k . When S k > 0, the control measure of this control point is effective, and S k the larger it is, the more significant the improvement effect on the system stability level; S k < 0, the control measure of this control point is invalid. At this time, the larger it is, the worse the influence on the system stability level.

[0014] Furthermore, in the step of determining the control sensitive point according to the sensitivity sorting result, calculate the control sensitivity of each alternative control point according to the sensitivity to generate a sorting result. The calculation formula of the control sensitivity of each alternative control point is as follows:

[0015] Wherein, represents the change in the power angle stability margin of the system after applying the k point control measure to the instability scenario; represents the control adjustment amount of the k th control point.

[0016] On the second aspect, the present invention also provides a system for determining the control sensitive point of power angle stability of a power system, including: A simulation scenario generation module, configured to perform time-domain simulation based on the operation mode and fault form of the power system to generate a basic scenario; A power angle stability sample set generation module, configured to collect the simulation data output by the basic scenario, calculate the power angle stability margin, judge the stability level of the system through the power angle stability margin. When the system is unstable, select the potential control points for generator tripping, apply control measures to form power angle stability samples, and the power angle stability samples are combined with the basic scenario to generate a power angle stability sample set; The power angle stability margin evaluation model generation module is used to take the time series curves of key electrical characteristics and control variables as input features, and the corresponding power angle stability margin as the output label, improve and train the CNN model to determine the optimal structure and parameters of the model, and generate the system power angle stability margin evaluation model; The sensitive point determination module is used to obtain the power angle stability margin improvement amount under different control measures based on the system power angle stability margin evaluation model in the case of power angle instability, calculate the sensitivity through the power angle stability margin improvement amount to screen out effective control points, and determine the control sensitive points according to the sensitivity sorting result, so as to complete the determination of the power angle stability control sensitive points of the power system.

[0017] In a third aspect, the present invention also provides a mobile terminal, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the method for determining the power angle stability control sensitive points of the power system as described above are implemented.

[0018] In a fourth aspect, the present invention also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method for determining the power angle stability control sensitive points of the power system as described above are implemented.

[0019] Compared with the prior art, the present invention has the following beneficial technical effects: The present invention provides a method for determining the power angle stability control sensitive points of a power system. By generating a basic scenario through time-domain simulation and collecting simulation data to calculate the power angle stability margin, the stable level of the power system can be accurately evaluated. Compared with traditional static or approximate methods, it can more realistically reflect the behavior of the power system in the dynamic process and improve the accuracy of stability evaluation. When the system becomes unstable, the present invention can select potential control points for generator tripping and apply control measures to form power angle stability samples. By constructing a power angle stability sample set and using the improved trained CNN model for power angle stability margin evaluation, the effects of different control measures can be accurately predicted and evaluated, which helps to screen out effective control measures and optimize them to improve the stability of the power system. The present invention calculates the power angle stability margin improvement amount under different control measures and uses sensitivity to screen out effective control points, and finally determines the control sensitive points. This method can accurately identify the control points that have the greatest impact on the stability of the power system, so as to apply control measures targeted and improve the control efficiency. Furthermore, the present invention constructs a quantization index for the system power angle stability margin based on EEAC. Considering different operating modes and fault forms of the power system, a data-driven system power angle stability margin evaluation model is proposed based on an improved CNN. A quantization relationship between the control measure quantity of the control point and the power angle stability level is established, providing an effective basis for the excavation of control sensitive points. The present invention defines an emergency control sensitivity index, uses the improved CNN model to judge the influence of the action of each control measure on the system power angle stability level, and screens out effective control measures; through the improved CNN regression prediction, the sensitivity ranking of effective control measures is realized. The proposed method can adapt to the power network with complex and changeable operating modes and provides technical support for the formulation of emergency control strategies for the power system. Brief Description of the Drawings

[0020] Figure 1 It is a flowchart of the method for determining the power angle stability control sensitive point in the embodiment of the present invention; Figure 2 It is the structure diagram of the improved CNN in the embodiment of the present invention; Figure 3 It is the flowchart of the power angle stability emergency control in the embodiment of the present invention; Figure 4 It is a schematic diagram of the deviation between the true value and the predicted value evaluated by different models in the embodiment of the present invention; Figure 5 It is a schematic diagram of the power angle swing curve of the machine group in northern Shaanxi in the embodiment of the present invention; Figure 6 It is the system power angle response curve after emergency control in the embodiment of the present invention; Figure 7 It is the schematic diagram of the principle of the system for determining the power angle stability control sensitive point in the embodiment of the present invention; In the figure: 1. Simulation scenario generation module; 2. Power angle stability sample set generation module; 3. Power angle stability margin evaluation model generation module; 4. Sensitive point determination module. Detailed Embodiment

[0021] In order to enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0022] The object of the present invention is to provide a method for determining a sensitive point of power angle stability control in a power system and related equipment, so as to solve the technical problem in the prior art of how to improve the accuracy of evaluating the enhancement effect of each controllable generator tripping resource on the power angle stability margin in a high-proportion renewable new energy power system under different operating modes.

[0023] The following further describes the present invention in detail with reference to the accompanying drawings: See Figure 1 , the present invention provides a method for determining a sensitive point of power angle stability control in a power system, including: Step 1, perform time-domain simulation based on the operating mode and fault form of the power system to generate a basic scenario; Step 2, collect the simulation data output by the basic scenario, and calculate the power angle stability margin. Judge the stability level of the system through the power angle stability margin. When the system is unstable, select potential generator tripping control points, apply control measures to form power angle stability samples, and the power angle stability samples are combined with the basic scenario to generate a power angle stability sample set; Specifically, the simulation data output by the basic scenario includes power angle time-series curves and key electrical parameter time-series information under different operating scenarios.

[0024] Among them, through power system time-domain simulation, power angle time-series curves and key electrical parameter time-series information under different operating scenarios are obtained, and the power angle stability of the system is judged by EEAC. EEAC realizes power angle stability discrimination based on the transient energy function. Its basic principle is to divide the multi-machine system into a critical machine group and a remaining machine group, and then equivalent the system to a two-machine system according to the inertia centers of the two machine groups, and further equivalent it to a single-machine infinite-bus system, and use the extended equal area criterion for stability analysis.

[0025] Based on the obtained power angle swing curve, identify coherent generators through the k-means clustering method to realize the division of the critical machine group C and the remaining machine group R, and use the local inertia center to replace the two machine groups respectively: (1) (2) In the formula: M C , , are respectively the equivalent inertia, equivalent power angle and equivalent speed of the critical machine group, M R , , are respectively the equivalent inertia, equivalent power angle and equivalent speed of the remaining machine group. This two-machine system can be further equivalent to a single-machine infinite-bus system as follows: (3) (4) (5) (6) (7) In the formula: , , , , are respectively the power angle, rotational speed, inertia, mechanical power and electromagnetic power parameters equivalent to the single-machine system. Based on the power angle characteristic curves before and after fault excision, calculate its accelerating area S A , decelerating area S B , then define the power angle stability margin of the system as follows: (8) , indicating that the decelerating area S B is greater than or equal to the accelerating area S A , the system is stable; , indicating that the decelerating area S B is less than the accelerating area S A , the system is unstable; The larger the [[ ]] represents the larger the difference between the decelerating area and the accelerating area of the system, and the larger the power angle stability margin of the system; When it is smaller, it indicates that the difference between the decelerating area and the accelerating area of the system is smaller, and the power angle stability margin of the system is smaller. According to the established transient stability margin index, quantitatively evaluate the stability level of each basic scenario. In particular, when the system is stable, due to the incomplete power angle characteristic curve, the decelerating area at this time specifically refers to the maximum decelerating area generated by the fitted complete power angle characteristic curve.

[0026] Based on the power angle stability margin quantitative evaluation index, conduct a stability analysis of the basic scenarios, and select the operating mode and fault combination with power angle instability and similar instability modes as the basic instability scenarios. Select potential control points for generator tripping in the critical machine group, apply different combinations of control measures to each instability scenario, generate a large amount of power angle stable and unstable sample data sets, and calculate their power angle stability margins using the EEAC index.

[0027] Step 3: Using the key electrical characteristic time series curve and control variables as input features and the corresponding power angle stability margin as the output label, perform improved training on the CNN model to determine the optimal structure and parameters of the model, and generate a system power angle stability margin evaluation model; Specifically, the traditional CNN structure is as Figure 2As shown in the figure, it includes an input layer, a convolutional layer, a downsampling layer (pooling layer), a fully connected layer, and an output layer. In the convolutional layer, different input features can be extracted through convolution kernel calculations; in the downsampling layer, that is, the pooling layer, local aggregation statistics are performed on the feature map output by the convolutional layer to achieve secondary feature extraction, further reducing the dimension of the feature map and ensuring a certain degree of scale invariance of the features; in the fully connected layer, by integrating local information of the data, key feature factors are extracted, and two-dimensional features are expanded into one-dimensional features; the output layer is a logistic regression layer, and according to the features output by the fully connected layer, the neuron nodes of the output layer classify them.

[0028] In order to improve the accuracy of the CNN model in extracting complex and high-dimensional features, the present invention optimizes the CNN network structure, adopts multiple convolutional layers to extract multi-dimensional features, appropriately reduces the size of the convolution kernel to reduce the number of calculation parameters, and improves the calculation speed.

[0029] At the same time, in the introduction of the non-linear link for optimization, the unsaturated non-linear function Leaky ReLU is used as the activation function, which can make the network better fit various features. The function expression is as follows: (9) This function can solve the saturation problem in the positive interval, make the output of some neurons be 0, reduce the dependence between parameters, and alleviate the gradient disappearance phenomenon. In addition, due to the simple operation, good sparsity, fast calculation speed and convergence time of the Leaky ReLU function, it does not need to rely on unsupervised pre-training, is more suitable for deep neural networks, can effectively avoid overfitting and improve the generalization ability of the model.

[0030] In addition, when the amount of data is too large, overfitting is likely to occur. Therefore, the loss function is regularized to correct the excessive extraction of feature variables and reduce the model complexity. This method can improve the generalization ability without overly weakening the fitting ability.

[0031] Since in engineering practice, misjudging ineffective control measures as effective may lead to the deterioration of the system stability level, when evaluating the prediction effect of the model, the effectiveness of emergency control actions should be concerned first, that is, whether the sensitivity prediction result is greater than 0. Here, the confusion matrix in statistics is introduced, as shown in Table 1.

[0032] Table 1 Confusion Matrix

[0033] (10) (11) (12) (13) In the formula, the accuracy rate A is the proportion of samples correctly predicted among the samples with the true value being "effective"; the recall rate R is the proportion of correctly judged results in the total observed values; the specificity S represents the proportion of samples correctly predicted among the samples with the true value being "invalid"; the misdiagnosis rate FPR is the proportion of samples with the true action being invalid but wrongly predicted as effective.

[0034] The accuracy rate A index characterizes the classification accuracy and is the main index for evaluating the overall prediction performance of the model. Since the main research is on the effective control action point, in addition to the accuracy rate A , the FPR representing the misjudgment rate of invalid actions should be focused on during evaluation.

[0035] To evaluate the actual prediction effect of the improved CNN model and prove the applicability of the model, when the improved CNN model is used for regression prediction, the root mean square error (RMSE) is used to calculate the sensitivity numerical prediction error as follows: (14) In the formula: N is the number of samples, y' ( i ) and y ( i ) are the predicted value and the true value of the i th sample respectively. RMSE The smaller T r is, the smaller the prediction error of the model is, and the better the prediction effect is. Therefore, based on RMSE, the prediction accuracy index (15) is defined to evaluate the prediction effect of the improved CNN model as follows:

[0036] Step 4: Based on the power angle stability margin evaluation model of the system, obtain the improvement amount of the power angle stability margin under different control measures when the power angle instability phenomenon occurs, calculate the sensitivity through the improvement amount of the power angle stability margin to screen out the effective control points, and determine the control sensitive points according to the sensitivity sorting result, thus completing the determination of the power angle stability control sensitive points of the power system. Specifically, when the change in the control input to the system state or stability index is small, it can be approximately considered that there is a linear correlation between the state or stability index and the control input. Based on the local linearization method, the correlation between the control measure and the change in transient stability margin is established as follows: (16) In the formula: represents the transient stability margin of the system after taking emergency control measures, represents the transient stability margin of the system without taking emergency control measures, represents the control adjustment amount of the k th control point. Therefore, the control sensitivity S k is defined as: (17) Then the sensitivity S k characterizes the influence effect of the control point k on the power angle stability margin of the system, representing the power angle stability margin changed by the unit control amount of the control point k . S k When S k >0, the control measure of this control point is effective, and the larger S k is, the more significant the improvement effect on the system stability level; When <0, the control measure of this control point is ineffective. At this time, the larger is, the more adverse the impact on the system stability level.

[0037] In formula (17), the power angle stability control sensitivity of each potential control point can be calculated by the following formula S k : (18) In the formula: represents the change in the power angle stability margin of the system after applying the k point control measure to the unstable scenario.

[0038] Based on the above-mentioned control sensitivity index, calculate the control sensitivity of each alternative control point according to formula (18), generate a sorting result, and the control point with the highest sensitivity is regarded as the most sensitive point with the highest sensitivity, and so on. Therefore, the online mining of power angle stability control sensitive points can be realized according to the sensitivity sorting result.

[0039] The specific process of the emergency control of the present invention is as shown in Figure 3As shown in the figure. In the off-line generation stage, the operation mode of the test system is adjusted, and various typical faults are simulated to obtain historical data or time-domain simulation data. The stability of the system is judged based on the EEAC power angle stability margin evaluation index, and combined alternative emergency control measures are applied to the unstable samples. The time-series information of the key electrical parameters of the system before and after the fault and the control variables are used as the sample input features, and the power angle stability margin is used as the output label to construct a sample set. The improved CNN model is trained to determine the model structure and parameters, and the model performance is evaluated using the test set. Finally, a power angle stability margin evaluation model with excellent prediction performance is generated.

[0040] In the on-line determination and application stage, when a fault occurs, the time-series values of the state variables are collected in real time by means of the on-line measurement system, and the existing stability margin evaluation index is used to determine whether the system is stable. If a power angle instability phenomenon occurs, based on the proposed model, the improvement amount of the power angle stability margin under different control measures is obtained, the effective control points are screened by calculating the sensitivity, and the control sensitive points are determined according to the sensitivity sorting result, and the emergency control actions are activated in turn. Finally, the transient stability of the system after applying the control is judged. If the system units still cannot restore synchronous stability, the data-driven emergency control will be executed again until stability is restored.

[0041] Example 1 The 750KV main network structure of the northern Shaanxi power grid in the Shaanxi power grid in 2025 is used to verify a method for determining the control sensitive points of power angle stability of a power system provided in this Example 1, where Figure 4 as shown.

[0042] Considering different operating conditions of the system, to ensure the comprehensiveness and practical application value of the samples, combined with the actual situation of the northern Shaanxi power grid, operating modes with new energy penetration rates of 50%, 60%, and 70% are set in this simulation system. At the same time, the system load level is set from 90% to 110%, increasing in steps of 5%. Under the corresponding operating modes, the actual power grid fault conditions are simulated, and different fault types are set, including 8 typical fault scenarios such as three-phase grounding short circuit, DC blocking, and complex AC-DC faults. Among them, the AC fault location is set at 10% to 90% of the AC line, increasing 9 fault points in steps of 10%. The fault duration is 0.1s - 0.25s, increasing in steps of 0.03s. The total simulation time is 10s, and a total of 5040 basic scenarios are generated by the simulation.

[0043] The K-means clustering method is used to identify the coherence of generator units. Based on the identification results, the generator units are grouped, and the system stability is judged according to the EEAC rule. In the basic scenario, 4635 stable samples and 905 unstable samples are obtained by calculating the stability margin through time-domain simulation. 16 groups of generators in the system are selected to participate in the coherence identification and the potential control capacity is provided. The corresponding control quantities are shown in Table 2 (benchmark capacity: 100 MW). For the case of system instability, the number of critical generator groups in different unstable samples is different. First, the generator tripping control points in the critical generator units are selected according to the generator unit grouping results. Then, considering the power supply demand of the system area and the diversity of generator tripping selection, different combined generator tripping control schemes are formed by tripping 1 generator unit and 2 generator units in the critical generator groups respectively. For the case of system stability, the control measure quantity is set to 0.

[0044] Table 2 Generator units participating in grouping and their controllable capacities

[0045] According to the above combinations of operating conditions and control measures, through the operating variables and control variables of the system, the corresponding power angle stability margin indexes before and after the system emergency control measures are calculated. The time series data of the power angle, speed, electromagnetic power, etc. of the multi-machine system under each operating scenario are collected as input features, and the power angle stability margin of the system is used as the output label. All samples are divided into a training set and a test set in a ratio of 4:1. Based on the generated training set samples, an improved CNN model is trained, and the prediction effect of the model is evaluated according to the test set.

[0046] In the test set, the deviation between the true value and the predicted value obtained by evaluating the proposed model is compared with traditional artificial intelligence methods, such as backpropagation (BP) neural network, decision tree (DT), and convolutional neural network (CNN), as Figure 3 shown. According to (13), the classification prediction results of control actions are shown in Table 3, and the prediction accuracies of different regression models are calculated, as shown in Table 4. The prediction results and the calculation results of accuracies intuitively show that the proposed method has better regression prediction effect, providing an effective guidance for the evaluation of control sensitivity.

[0047] Table 3 Classification prediction results of control actions

[0048] Table 4 Prediction accuracies of different regression models

[0049] During online application, for a non-specific scenario of similar power angle instability modes, the control sensitive points during instability can be predicted through the above model. Under the operating mode with a new energy penetration rate of 70% and a load level of 100% in northern Shaanxi, when an N-1 fault occurs in the power grid, the fault location is at the 10% position of the 750 kV line from Shuofang to Yuheng, and the fault clearing time is 0.15 s after the fault occurs, the power angle response curve in northern Shaanxi is as Figure 5 shown. The results show that the units in northern Shaanxi are separated into two groups, presenting an instability phenomenon and unable to return to a stable state. Through the K-means method, the critical machine groups are identified, and the alternative generator tripping points include Shanyu**, Shaanxi**, Shanyang**, Shanqing**, Shandian**, Shanjie**, Shanqing**, Shanqing**, Shanshuang**. Based on the EEAC stability margin index, the power angle stability margin of the system is calculated to be -0.2558, and the system is unstable.

[0050] At 0.25 s after the fault occurs, the stability control measures are executed. The improved CNN model proposed is used to predict the improvement degree of the system stability margin after applying different control measures, and then the control sensitivity of each control point in the system is solved. The prediction results are shown in Table 5.

[0051] Table 5 Increase in system stability level after each alternative control measure

[0052] The results in Table 5 show that removing generators at different control points has different effects on improving the system stability margin. According to the sensitivity calculation formula in (12), the predicted emergency control sensitivities of each control point are shown in Table 6.

[0053] Table 6 Predicted sensitivities of generator tripping control points

[0054] By comparing with the true values, the improved CNN network more accurately fits the sensitivity values, and the sorting is consistent with the true values. Therefore, for this instability mode, the power angle stability control sensitive points of the system and the implementation control sensitivity levels are sorted from high to low as Shanyu**, Shandian**, Shanqing**, Shaanxi**, Shanjie**, Shanyang**, Shanqing**, Shanqing**, Shanshuang**. Among them, the generator tripping sensitivity of the Shanyu** node is the highest, indicating that when the same control amount of stability control measures are implemented at each control point, removing the units at the Shanyu** control node has the best effect on restoring the system stability level.

[0055] Generator tripping operations are sequentially performed on each control node according to the sensitivity magnitude until the system returns to stability. The results show that after sequentially removing the units of Shanyu**, Shandian**, Shanqing**, and Shaanxi** with a removal ratio of 100%, the system returns to stability, and the power angle response curve is as Figure 6As shown in the figure. Therefore, the proposed method can accurately predict the sensitivity of emergency control generator tripping actions, judge effective generator tripping actions, realize the excavation of sensitive points for power angle stability control of the power system, and provide a basis for optimizing emergency control decisions.

[0056] In summary, a method for determining sensitive points for power angle stability control of a power system provided in this embodiment generates a basic scenario through time-domain simulation and collects simulation data to calculate the power angle stability margin, thereby being able to accurately evaluate the stability level of the power system. Compared with traditional static or approximate methods, it can more realistically reflect the behavior of the power system during the dynamic process and improve the accuracy of stability assessment. When the system becomes unstable, the present invention can select potential generator tripping control points and apply control measures to form power angle stability samples. By constructing a power angle stability sample set and using an improved trained CNN model to evaluate the power angle stability margin, it is possible to accurately predict and evaluate the effects of different control measures, which helps to screen out effective control measures and optimize them to improve the stability of the power system. The present invention calculates the increase in the power angle stability margin under different control measures and uses sensitivity to screen effective control points, and finally determines the control sensitive points. This method can accurately identify the control points that have the greatest impact on the stability of the power system, thereby applying control measures targeted to improve control efficiency. Embodiment 2 According to Figure 7 As shown in the figure, this embodiment provides a system for determining sensitive points for power angle stability control of a power system, including: A simulation scenario generation module 1, configured to perform time-domain simulation based on the operation mode and fault form of the power system to generate a basic scenario; A power angle stability sample set generation module 2, configured to collect the simulation data output by the basic scenario, calculate the power angle stability margin, judge the stability level of the system through the power angle stability margin, select potential generator tripping control points when the system is unstable, apply control measures to form power angle stability samples, and generate a power angle stability sample set in combination with the basic scenario; A power angle stability margin evaluation model generation module 3, configured to use the key electrical feature time series curve and control variables as input features and the corresponding power angle stability margin as the output label to improve the training of the CNN model to determine the optimal structure and parameters of the model, and generate a system power angle stability margin evaluation model; A sensitive point determination module 4, configured to obtain the increase in the power angle stability margin under different control measures based on the system power angle stability margin evaluation model in the case of power angle instability, calculate the sensitivity through the increase in the power angle stability margin to screen effective control points, and determine the control sensitive points according to the sensitivity ranking result, completing the determination of sensitive points for power angle stability control of the power system.

[0057] Embodiment 3 The present invention also provides a mobile terminal, including a memory, a processor, and a computer program stored in the memory and executable on the processor, such as a program for determining the sensitive points of power angle stability control in a power system.

[0058] When the processor executes the computer program, the steps of the above method for determining the sensitive points of power angle stability control in a power system are implemented, for example: Performing time-domain simulation based on the operation mode and fault form of the power system to generate a basic scenario; Collecting the simulation data output by the basic scenario, calculating the power angle stability margin, judging the stability level of the system through the power angle stability margin. When the system is unstable, selecting potential generator tripping control points, applying control measures to form power angle stability samples, and combining the power angle stability samples with the basic scenario to generate a power angle stability sample set; Using the key electrical feature time-series curve and control variables as input features, and the corresponding power angle stability margin as the output label, improving and training the CNN model to determine the optimal structure and parameters of the model, and generating a power angle stability margin evaluation model for the system; Based on the power angle stability margin evaluation model for the system, obtaining the power angle stability margin improvement amounts under different control measures in the case of power angle instability, calculating the sensitivity through the power angle stability margin improvement amounts to screen out effective control points, and determining the control sensitive points according to the sensitivity ranking results, thus completing the determination of the sensitive points of power angle stability control in the power system. Alternatively, when the processor executes the computer program, the functions of each module in the above system are implemented, for example: A simulation scenario generation module 1, configured to perform time-domain simulation based on the operation mode and fault form of the power system to generate a basic scenario; A power angle stability sample set generation module 2, configured to collect the simulation data output by the basic scenario, calculate the power angle stability margin, judge the stability level of the system through the power angle stability margin. When the system is unstable, select potential generator tripping control points, apply control measures to form power angle stability samples, and combine the power angle stability samples with the basic scenario to generate a power angle stability sample set; A power angle stability margin evaluation model generation module 3, configured to use the key electrical feature time-series curve and control variables as input features, and the corresponding power angle stability margin as the output label, improve and train the CNN model to determine the optimal structure and parameters of the model, and generate a power angle stability margin evaluation model for the system; The sensitive point determination module 4 is configured to obtain the improvement amount of the power angle stability margin under different control measures based on the power angle stability margin evaluation model of the system in the event of power angle instability, calculate the sensitivity through the improvement amount of the power angle stability margin to screen out effective control points, and determine the control sensitive points according to the sensitivity sorting result, thereby completing the determination of the power angle stability control sensitive points of the power system.

[0059] For example, the computer program can be divided into a simulation scenario generation module 1, a power angle stability sample set generation module 2, a power angle stability margin evaluation model generation module 3, and a sensitive point determination module 4; The specific functions of each module are as follows: The simulation scenario generation module 1 is configured to perform time-domain simulation based on the operation mode and fault form of the power system to generate a basic scenario; The power angle stability sample set generation module 2 is configured to collect the simulation data output by the basic scenario, calculate the power angle stability margin, judge the stability level of the system through the power angle stability margin. When the system is unstable, select potential control points for generator tripping, apply control measures to form power angle stability samples, and generate a power angle stability sample set by combining the power angle stability samples with the basic scenario; The power angle stability margin evaluation model generation module 3 is configured to use the key electrical characteristic time series curve and control variables as input features, and the corresponding power angle stability margin as the output label to improve and train the CNN model to determine the optimal structure and parameters of the model, and generate a system power angle stability margin evaluation model; The sensitive point determination module 4 is configured to obtain the improvement amount of the power angle stability margin under different control measures based on the power angle stability margin evaluation model of the system in the event of power angle instability, calculate the sensitivity through the improvement amount of the power angle stability margin to screen out effective control points, and determine the control sensitive points according to the sensitivity sorting result, thereby completing the determination of the power angle stability control sensitive points of the power system.

[0060] The mobile terminal may be a computing device such as a desktop computer, a notebook, a palm computer, and a cloud server. The mobile terminal may include, but is not limited to, a processor and a memory.

[0061] The processor may be a Central Processing Unit (CPU), or may also be other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field-Programmable Gate Arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the mobile terminal, connecting various parts of the entire mobile terminal through various interfaces and circuits.

[0062] The memory can be used to store the computer programs and / or modules. The processor realizes various functions of the mobile terminal by running or executing the computer programs and / or modules stored in the memory, and by calling the data stored in the memory.

[0063] The memory mainly includes a program storage area and a data storage area. Among them, the program storage area can store an operating system, application programs required for at least one function (such as a sound playback function, an image playback function, etc.); the data storage area can store data created according to the use of the mobile phone (such as audio data, phone book, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, at least one magnetic disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0064] Embodiment 4 The present invention also provides a computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, the steps of the method for determining a power angle stability control sensitive point of a power system are implemented.

[0065] If the modules / units integrated in the mobile terminal are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium.

[0066] Based on such understanding, all or part of the processes in the above method of the present invention can also be completed by instructing relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, the steps of the above aggregation reinforcement learning resource scheduling method can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file or some intermediate form, etc.

[0067] The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disc, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electrical carrier signal, telecommunication signal, and software distribution medium, etc.

[0068] It should be noted that the content included in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electrical carrier signals and telecommunication signals.

[0069] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications or equivalent replacements can still be made to the specific implementation manners of the present invention, and any modification or equivalent replacement that does not depart from the spirit and scope of the present invention shall be covered by the protection scope of the claims of the present invention.

Claims

1. A method for determining a sensitive point of power angle stability control in a power system, characterized in that, Including: Performing time-domain simulation based on the operation mode and fault form of the power system to generate a basic scenario; Collecting the simulation data output by the basic scenario, calculating the power angle stability margin, judging the stability level of the system through the power angle stability margin. When the system becomes unstable, selecting potential control points for generator tripping, applying control measures to form power angle stability samples, and generating a power angle stability sample set by combining the power angle stability samples with the basic scenario; Using the key electrical feature time-series curve and control variables as input features, and the corresponding power angle stability margin as the output label, improving and training the CNN model to determine the optimal structure and parameters of the model, and generating a system power angle stability margin evaluation model; Based on the system power angle stability margin evaluation model, obtaining the improvement amount of the power angle stability margin under different control measures in the case of power angle instability, calculating the sensitivity through the improvement amount of the power angle stability margin to screen effective control points, and determining the control sensitive points according to the sensitivity sorting result, thus completing the determination of the power angle stability control sensitive points of the power system.

2. The method for determining a sensitive point for power angle stability control of a power system according to claim 1, characterized in that The simulation data output by the basic scenario includes the power angle time-series curve and the key electrical parameter time-series information under different operation scenarios.

3. The method for determining a sensitive point for power angle stability control of a power system according to claim 1, characterized in that, In the step of judging the stability level of the system through the power angle stability margin, the extended equal area criterion is used for stability analysis, including: Dividing the multi-machine system into a critical machine group and a remaining machine group, and then equivalenting the system to a two-machine system, i.e., a single-machine infinite bus system, according to the inertia centers of the critical machine group and the remaining machine group; Calculate the accelerating area of the power angle time series curve based on the power angle time series curves before and after fault clearing S A and the decelerating area S B Then, through the said accelerating area S A and the decelerating area S B calculate the power angle stability margin of the system, and the formula is as follows: Among them, represents the deceleration area S B is greater than or equal to the acceleration area S A , the system is stable; represents the deceleration area S B is less than the acceleration area S A , the system is unstable; The larger it is, the greater the difference between the deceleration area and the acceleration area of the system, and the greater the system stability margin; The smaller it is, the smaller the difference between the deceleration area and the acceleration area of the system, and the smaller the system stability margin.

4. The method for determining the sensitive point of power angle stability control of a power system according to claim 1, characterized in that The CNN model includes an input layer, a convolutional layer, a downsampling layer, a fully connected layer, and an output layer; The process of improving the CNN model is as follows: Adopting multiple convolutional layers to extract multi-dimensional features and reducing the size of the convolutional kernel to reduce the number of calculation parameters; Using the unsaturated non-linear function Leaky ReLU as the activation function to fit various features; Regularizing the loss function to correct the over-extraction of feature variables and reduce the model complexity.

5. The method for determining the sensitive point of power angle stability control of a power system according to claim 1, characterized in that, Performing a predictive evaluation on the generated system power angle stability margin evaluation model, calculating the sensitivity numerical prediction error using the root mean square error, and the calculation formula is as follows: Among them, N is the number of samples, y' ( i ) and y ( i ) are the predicted value and the true value of the i -th sample, respectively.

6. The method for determining the sensitive point of power angle stability control of a power system according to claim 1, characterized in that In the step of calculating the sensitivity through the improvement amount of the power angle stability margin to screen effective control points, the process of determining the sensitivity is as follows: Establishing the correlation relationship between the control measure and the change amount of the transient stability margin, and the expression formula is as follows: ； Wherein: represents the transient stability margin of the system after taking emergency control measures, represents the transient stability margin of the system without taking emergency control measures, represents the k control adjustment amount of the Determining the sensitivity based on the correlation relationship between the control measure and the change amount of the transient stability margin, and the calculation formula is as follows: ； Among them, represents the control adjustment amount of the k th control point; represents the correlation between the control measure and the change in transient stability margin; S k characterizes the influence of the control point k on the power angle stability margin of the system, representing the power angle stability margin changed by the unit control amount of the control point k ; when S k > 0, the control measure of this control point is effective, and S k the larger it is, the more significant the improvement effect on the system stability level; S k < 0, the control measure of this control point is ineffective, and at this time the larger it is, the more adverse the impact on the system stability level.

7. The method for determining the sensitive point of power angle stability control of a power system according to claim 6, characterized in that, In the step of determining the control sensitive points according to the sensitivity sorting result, calculating the control sensitivity of each alternative control point according to the sensitivity to generate a sorting result, and the calculation formula of the control sensitivity of each alternative control point is as follows: Among them, represents the change in the power angle stability margin of the system after applying k point control measures ; represents the control adjustment amount of the k th control point.

8. A system for determining sensitive points of power angle stability control in a power system, characterized in that, Including: A simulation scenario generation module for performing time-domain simulation based on the operation mode and fault form of the power system to generate a basic scenario; The power angle stability sample set generation module is used to collect the simulation data output by the basic scenario, calculate the power angle stability margin, judge the stability level of the system through the power angle stability margin. When the system is unstable, select the potential control points for generator tripping, apply control measures to form power angle stability samples, and the power angle stability samples are combined with the basic scenario to generate a power angle stability sample set; The power angle stability margin evaluation model generation module is used to take the key electrical feature time series curve and control variables as input features, and the corresponding power angle stability margin as the output label, improve and train the CNN model to determine the optimal structure and parameters of the model, and generate a system power angle stability margin evaluation model; The sensitive point determination module is used to obtain the power angle stability margin improvement amount under different control measures based on the system power angle stability margin evaluation model in the case of power angle instability, calculate the sensitivity through the power angle stability margin improvement amount to screen effective control points, and determine the control sensitive points according to the sensitivity sorting result, completing the determination of the power angle stability control sensitive points of the power system.

9. A mobile terminal, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method for determining the power angle stability control sensitive points of the power system according to any one of claims 1-7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the method for determining the power angle stability control sensitive points of the power system according to any one of claims 1-7.

Images