A transient power angle stability online evaluation method, device and system

Abstract

The application discloses a kind of transient power angle stability online evaluation method, device and system, belong to power system stability analysis technical field, by prediction model, the relative speed of the disturbed serious generator set pair in power system is predicted, and then relative power angle is predicted;Relative speed and relative power angle are used to carry out PMLE estimation and PMLE phase plane trajectory calculation, and finally according to PMLE and its phase plane trajectory, the stability of power system transient power angle is quickly and accurately evaluated.Utilize the model-free method of PMLE to realize the online fast evaluation method of transient stability, can greatly reduce the time-consuming while maintaining excellent accuracy to all kinds of power angle swing;Thus solve the technical problem of low rate of existing transient power angle stability evaluation method.In addition, the speed and power angle of each generator, without involving system structure and other model parameters, so data is easy to obtain, so that the method application is more in line with the actual, can be applied to engineering practice.

Description

Technical Field

[0001] This invention belongs to the field of power system stability analysis technology, and more specifically, relates to an online assessment method, device and system for transient power angle stability. Background Technology

[0002] The intermittent energy integration into large-scale power systems has exacerbated transient stability problems. Synchronous machine out-of-sync issues can easily trigger widespread blackouts, resulting in severe losses. Real-time and accurate transient stability assessment (TSA) is crucial for ensuring the safety of power systems.

[0003] Benefiting from the development of synchronizer technology and wide-area measurement technology, large-scale power systems can acquire data and perform assessments in real time. To date, model-driven and data-driven transient stability assessment methods have been developed. Model-driven methods mainly include time-domain simulation analysis, the equal-area criterion method, and the transient energy equation method, but all have some drawbacks. Time-domain simulation analysis has high computational costs, making it difficult to achieve ultra-real-time computing. Limited by the need for accurate model information, model-driven methods struggle to guarantee both speed and accuracy in large-scale power systems.

[0004] Data-driven Transient Stability Assessment (TSA) methods effectively avoid the problem of inaccurate physical models and are gradually becoming the mainstream trend. Support Vector Machines, Decision Trees, and Extreme Learning Machines have all achieved good results in TSA. However, this involves a large amount of sample acquisition and agent training time. In addition, the generalization and transferability across different scenarios still pose a significant challenge. The model-free Maximum Lyapunov exponent (MLE) method can utilize the general characteristics of transient stability in power systems, thus effectively solving the problem of poor transferability and eliminating the need for agent training. However, the fixed observation window of MLE inevitably leads to significant time consumption in order to ensure accuracy, making it impossible to quickly assess the stability of transient power angles. Summary of the Invention

[0005] To address the aforementioned deficiencies or improvement needs of existing technologies, this invention provides an online method, apparatus, and system for assessing transient power angle stability. Its purpose is to predict the relative rotational speed of severely disturbed generator pairs in a power system using a predictive model, thereby predicting the relative power angle; to estimate the power phase angle (PMLE) and calculate the PMLE phase plane trajectory using the relative rotational speed and relative power angle; and finally, to quickly and accurately assess the stability of the power system's transient power angle based on the PMLE and its phase plane trajectory, thus solving the technical problem of low speed in existing transient power angle stability assessment methods.

[0006] To achieve the above objectives, according to one aspect of the present invention, a method for online evaluation of transient power angle stability is provided, comprising:

[0007] S1: After the fault is cleared, the relative speed of the severely disturbed generator pair is predicted using a prediction model with a variable regression equation.

[0008] S2: Input the relative power angle and relative speed collected from the severely disturbed generator set pair into the self-memory prediction model to obtain the predicted relative power angle;

[0009] S3: Use the relative rotational speed and the relative power angle to determine the oscillation mode classification, and then obtain the corresponding maximum Lyapunov MLE time-domain curve, denoted as PMLE time-domain curve;

[0010] S4: Calculate the differential value of PMLE using the discrete data corresponding to the PMLE time-domain curve to establish the phase plane trajectory of PMLE;

[0011] S5: Analyze the PMLE time-domain curve and the PMLE phase plane trajectory to evaluate the stability of the transient power angle.

[0012] In one embodiment, in S1:

[0013] The rotational speed prediction includes: prediction model I based on an exponential function regression model and prediction model II based on a Fourier function prediction model;

[0014] After the fault is cleared, prediction model I is used until the relative speed increases to cross zero, at which point prediction model II is switched.

[0015] In one embodiment, the prediction model I is represented as: Where r is the order of the regression equation, γ m ω represents the regression coefficient of the m-th order, t represents the current time, and ω represents the relative rotational speed.

[0016] In one embodiment, the prediction model II is represented as:

[0017]

[0018] Where r is the order of the regression equation, a m b m Let f be the two regression coefficients of the m-th order. u Let t be the frequency of the periodic oscillation, and t be the current time.

[0019] In one embodiment, S5 includes:

[0020] If the PMLE curve of the severely disturbed generator pair initially decreases, during the first oscillation, the parameter A corresponding to the inflection point of the phase plane trajectory and at some subsequent moment will be... λ If the value is greater than 1, the severely disturbed generator set is assessed as having multiple swing instability.

[0021] If the relative angle of attack is considered as a cosine function, then A λ The parameters of the fitted function are obtained by... The calculation is obtained, where t is the current time and λ is the PMLE value. For the parameters of the fitted function, w λ The angular velocity w is the periodic oscillation of the relative work angle. λ =2πf u ;f u The frequency of the periodic oscillation.

[0022] In one embodiment, A λ The expression: A is the amplitude of the relative work angle fitted cosine curve, t m δ is the time interval between the original trajectory and neighboring trajectories determined by PMLE estimation. n and δ m(n) These are the initial points of the original trajectory and the neighboring trajectories, respectively.

[0023] In one embodiment, S5 includes:

[0024] If the PMLE curve of the severely disturbed generator pair initially decreases, during the first oscillation, the intercept D of the phase plane trajectory on the PMLE axis at the inflection point and at some later time... PMLE If the value is less than 0, the corresponding severely disturbed generator set is assessed as stable.

[0025] In one embodiment, prior to S1, the method further includes:

[0026] After a fault occurs, the rotational speed of each generator unit is collected in real time.

[0027] At the fault clearing time, calculate the absolute value of the difference between the speed of each generator set and the speed of the preset generator set, and use a preset multiple of the maximum absolute value of the difference as the threshold.

[0028] The absolute value of the difference corresponding to each generator set is compared with the threshold. If it is greater than the threshold, the generator set and the preset generator set are determined to be the severely disturbed generator set pair.

[0029] The preset generator set is the generator set with the smallest speed change.

[0030] According to another aspect of the present invention, a transient power angle stability online evaluation device is provided, comprising:

[0031] The first prediction module is used to predict the relative speed of severely disturbed generator pairs using a prediction model with a variable regression equation after the fault is cleared.

[0032] The second prediction module is used to input the relative power angle and relative speed collected from the severely disturbed generator set into the self-memory prediction model to obtain the predicted relative power angle.

[0033] The acquisition module is used to determine the oscillation mode classification using the relative rotational speed and the relative power angle, and then obtain the corresponding maximum Lyapunov MLE time-domain curve, denoted as PMLE time-domain curve;

[0034] A module is established to calculate the differential value of the PMLE using the discrete data corresponding to the PMLE time-domain curve, so as to establish the phase plane trajectory of the PMLE;

[0035] The evaluation module is used to analyze the PMLE time-domain curve and the PMLE phase plane trajectory to evaluate the stability of the transient power angle.

[0036] According to another aspect of the present invention, a transient power angle stability online evaluation system is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described transient power angle stability online evaluation method.

[0037] According to another aspect of the present invention, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described online evaluation method for transient power angle stability.

[0038] In summary, compared with the prior art, the above-described technical solutions conceived by this invention can achieve the following beneficial effects:

[0039] (1) This scheme predicts the relative speed of severely disturbed generator pairs in the power system using a predictive model, and then predicts the relative power angle. It uses the relative speed and relative power angle to estimate the power system's transient power angle and calculate the PMLE phase plane trajectory. Finally, it rapidly and accurately assesses the stability of the power system's transient power angle based on the PMLE and its phase plane trajectory. Utilizing a model-free method for PMLE to achieve rapid online assessment of transient stability significantly reduces time consumption while maintaining excellent accuracy for all types of power angle fluctuations; thus solving the technical problem of low speed in existing transient power angle stability assessment methods. Furthermore, the speed and power angle of each generator do not involve model parameters such as system structure, making the data readily available and enabling the method to be applied more realistically in engineering practice.

[0040] (2) This scheme is set to start using the prediction model I after the fault is cleared, and then switch to the prediction model II when the relative speed rises to the zero point. The beneficial effect is that it can maintain a high prediction accuracy for the prediction of the relative speed curves of the first swing instability, multiple swing instability and multiple swing stability, and improve the reliable prediction window length.

[0041] (3) In this scheme, prediction model I is represented as: The beneficial effect is that it can reflect the rapid divergence characteristics of the first swing instability, but will not cause misjudgment.

[0042] (4) This scheme expresses prediction model II as: The beneficial effect is that it can accurately reflect the periodic oscillation characteristics.

[0043] (5) This scheme utilizes the parameter A corresponding to the phase plane trajectory at the inflection point and at a certain time thereafter. λ As a basis for stability assessment, if A λ If the value is greater than 1, the severely disturbed generator set is assessed as having multiple swing instability; the beneficial effect is that the judgment index is simple and easy to obtain, and the judgment result is accurate.

[0044] (6) This plan will include A λ The expression is set as: The beneficial effect is that it simplifies the calculation formula into an index while maintaining the same accuracy.

[0045] (7) If the PMLE curve of the severely disturbed generator pair in this scheme initially decreases, during the first oscillation, the intercept D of the phase plane trajectory on the PMLE axis at the inflection point and at some point thereafter. PMLE If the value is less than 0, the corresponding severely disturbed generator set is assessed as stable; the benefit is that it allows for accurate and rapid assessment of stability.

[0046] (8) This scheme calculates the absolute value of the difference between the speed of each generator set and the speed of the preset generator set, and uses the preset multiple of the maximum absolute value of the difference as the threshold; compares the absolute value of the difference corresponding to each generator set with the threshold to determine the severely disturbed generator set pair; the beneficial effect is: screening key generator set pairs for evaluation and analysis, greatly reducing the amount of calculation. Attached Figure Description

[0047] Figure 1 The flowchart illustrates an online evaluation method for transient power angle stability provided in this embodiment of the invention.

[0048] Figure 2 This is a structural diagram of a wind power testing system provided in an embodiment of the present invention.

[0049] Figure 3 This is a schematic diagram of a variable prediction model for relative rotational speed provided in an embodiment of the present invention.

[0050] Figure 4 The PMLE phase plane trajectory classification diagram provided in the embodiments of the present invention.

[0051] Figure 5a , Figure 5b and Figure 5c The figure shows the results of a typical instability calculation example provided in the embodiments of the present invention.

[0052] Figure 6a , Figure 6b and Figure 6c The figure shows the results of a typical stable calculation example provided in the embodiments of the present invention.

[0053] Figure 7a and Figure 7b A comparison chart of evaluation time consumption for different fault locations and clearing times provided for embodiments of the present invention.

[0054] Figure 8 A comparison chart of the accuracy and time consumption of different prediction windows provided in embodiments of the present invention. Detailed Implementation

[0055] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0056] The online evaluation method for transient power angle stability of power systems provided by this invention is as follows: Figure 1 As shown. The specific embodiment is an improved IEEE New England 39-bus system, comprising 10 synchronous generators, with 200MW doubly-fed wind farms connected to buses 3, 4, 16, 17, and 18 respectively, as follows. Figure 2 As shown. The load power is 6150.5MW, and the wind power penetration rate is 16.26%. The DFIG model includes generator-side and grid-side converter control. The generator-side converter adopts MPPT control. The doubly-fed wind farm adopts multiple DFIG aggregated equivalents with a rated power of 2MW each. The load adopts a constant impedance model. The PMU sampling rate for collecting data and calculating PMLE is 100 times / second.

[0057] First, using two typical faults as examples with the wind turbine not in operation, the specific steps of this invention for evaluating the IEEE standard system are explained. Finally, the applicability of this invention to wind power systems is tested with the wind turbine in operation. The two typical faults are: 1) A three-phase short circuit fault occurs at t=1s near busbar 26 on line 26-29. c = Disconnected at 1.25s; 2) Taking a three-phase short circuit fault occurring at point 17 near busbar 17 on line 17-18 at t=1s as an example, line 17-18 at t c Disconnect at 1.25s.

[0058] See Figure 1 The present invention provides a method for online evaluation of transient power angle stability, which includes the following steps: S1: After the fault is cleared, the relative speed of the severely disturbed generator pair is predicted using a prediction model with a variable regression equation;

[0059] S2: Input the relative power angle and relative speed collected from the severely disturbed generator pair into the self-memory prediction model to obtain the predicted relative power angle;

[0060] S3: Use relative speed and relative power angle to determine the oscillation mode classification, and then obtain the corresponding maximum Lyapunov MLE time domain curve, denoted as PMLE time domain curve;

[0061] S4: Calculate the differential value of PMLE using the discrete data corresponding to the PMLE time-domain curve, so as to establish the phase plane trajectory of PMLE;

[0062] S5: Analyze the PMLE time-domain curve and the PMLE phase plane trajectory to evaluate the stability of the transient power angle.

[0063] In one embodiment, the evaluation begins first, triggered by a fault detection signal. After a fault occurs, the generator power angle and speed are acquired and estimated in real time using a phasor measurement unit. At the fault clearing time, each generator unit is sorted according to the absolute value of its speed relative to the unit with the smallest change, and pairs of generator units with a relative speed greater than 0.7 of the maximum relative speed are selected as the severely disturbed generator unit pairs.

[0064] Specifically, for fault 1, the severely disturbed generator sets at the time of fault clearing are G9-G10. For fault 2, the severely disturbed generator sets are G7-G10.

[0065] In one embodiment, S1, relative speed prediction. The relative speed of the severely disturbed generator pair is predicted using a variable regression equation model with different prediction step sizes. When a specific moment is reached, the prediction model changes from I to II.

[0066] Specifically, such as Figure 3As shown, prediction model I is an exponential function regression model. In the initial period after fault clearance, time-series data is limited; therefore, a shorter observation window can reliably predict the rotational speed within time window W1. The expression for prediction model I is as follows:

[0067]

[0068] Where, γ m Here, r represents the regression coefficient, and r is the order of the regression equation. In this example, the order is chosen to be 3.

[0069] Prediction Model II is a Fourier function prediction model. With increased actual time series data, a longer observation window can be selected, and the prediction time window W2 can be trusted. The expression for Prediction Model II is as follows:

[0070]

[0071] Among them, a m b m f is the regression coefficient. u The frequency of the periodic oscillation.

[0072] Model switching: After the fault is cleared, Model I sliding prediction is used until the speed increases at the zero point, then Model II is switched.

[0073] For fault 1, the speed prediction model remains Model I, with a data observation window of 10 sampling points and a prediction time window W1 of 30ms. The speed prediction results are as follows: Figure 5a As shown; for fault 2, the relative speed crosses 0 at t=2.16s, so the prediction model is switched to II, the data observation time window is 10 sampling points, the prediction time window W1 is 30ms, after the model switch the data observation time window is 30 sampling points, and the prediction time window W2 is 40ms.

[0074] In one embodiment, S2 is the relative power angle prediction. Based on the collected relative power angle and the relative rotational speed predicted in S1, the predicted power angle is output after passing through a self-memory prediction model.

[0075] Specifically, in step S2, the power angle prediction employs a self-memory model, integrating the system's dynamic differential equation mechanism and historical data. The discrete self-memory equation is:

[0076]

[0077]

[0078] Where p is the order of the self-memory equation; α i θ i These are self-memory parameters.

[0079] By obtaining self-memory parameters through the least squares method, the relative power angle at step k can be predicted using this model, which utilizes the predicted relative rotational speed and the power angle predicted before step k.

[0080]

[0081] The predicted power angles for faults 1 and 2 are as follows: Figure 5b and Figure 6a As shown. This invention employs a self-memory prediction using a variable regression model that fits different actual trends, and utilizes the trend information contained in the PMLE phase plane, making the evaluation method applicable to various trends in the angle of motion, and significantly improving its speed.

[0082] In one embodiment, S3, PMLE sliding estimation. The oscillation mode classification and two parameters are determined by the relative rotational speed predicted in S1 and the relative power angle predicted in S2, and the time-domain curve of PMLE is continuously estimated by sliding using recursive least squares method.

[0083] Specifically, in step S3, the sliding estimate of PMLE is obtained through recursive least squares. The PMLE estimation equation is:

[0084]

[0085]

[0086] in, For the estimated PMLE, and These are the original trajectory point and the adjacent trajectory point at step k, respectively.

[0087] Based on the predicted power angle of S2, for fault 1, the PMLE can be estimated starting at t = 1.35s; for fault 2, the PMLE can be estimated starting at t = 1.73s. The estimated PMLEs for faults 1 and 2 are as follows: Figure 5c and Figure 6b As shown.

[0088] In one embodiment, S4, the phase plane trajectory of the PMLE is calculated. The differential value of the PMLE obtained in S3 is approximated using discrete difference data, and the phase plane trajectory of the PMLE is established. The trajectory of the PMLE in the second quadrant of the phase plane corresponds to the first oscillation process of the PMLE.

[0089] Specifically, such as Figure 4As shown, in step S4, the phase plane of PMLE is the plane composed of PMLE and its derivative. During the initial swing phase of PMLE, and when the phase plane trajectory has a negative slope, the slope of the trajectory tangent gradually decreases before the inflection point (if any), while the intercept on the PMLE axis gradually increases. After the inflection point, the slope gradually increases, while the intercept on the PMLE axis gradually decreases. The PMLE time-domain trajectory after the inflection point is approximately as follows:

[0090]

[0091] Where λ is PMLE, A λ , Let w be the parameters of the fitted function. λ =2πf u .

[0092] The phase plane trajectory of PMLE calculated for fault 2 is as follows: Figure 6c As shown.

[0093] In one embodiment, S5, stability assessment. Based on S3 and S4, transient power angle stability is assessed according to the stability criteria of PMLE and its phase plane trajectory.

[0094] Specifically, the stability evaluation criteria in step S5 are as follows:

[0095] Criterion 1: If the PMLE curve of a severely disturbed generator pair rises initially, then the severely disturbed generator pair is assessed as having initial swing instability.

[0096] Criterion 2: If the PMLE curve of a severely disturbed generator pair initially decreases, and during the first oscillation, the phase plane trajectory at the inflection point and at some subsequent moment, parameter A... λ If the value is greater than 1, the severely disturbed generator set is assessed as having multiple swing instability.

[0097] Criterion 3: If the PMLE curve of a severely disturbed generator pair initially decreases, during the first oscillation, the intercept D of the phase plane trajectory on the PMLE axis at the inflection point and at some subsequent moment... PMLE If the value is less than 0, the severely disturbed generator set is assessed as stable.

[0098] Criterion 4: For the power system as a whole, if all severely disturbed generator pairs are stable, then the system is stable; if any severely disturbed generator pair experiences initial swing instability or multiple swing instability, then the system as a whole is unstable.

[0099] For fault 1, at t = 1.61s, it was assessed as unstable because PMLE > 0, and this data was obtained from the prediction at t = 1.35s; therefore, the system was assessed as unstable, taking 0.10s. For fault 2, at t = 2.72s, due to D...PMLE =-0.19<0 is evaluated as stable, and this data was obtained from the prediction at t=2.32s, therefore the system is evaluated as stable, taking 1.07s.

[0100] Taking fault 1 as an example, the test results of this invention are presented, and the method proposed in this invention is compared with other methods. The test results using the autoregressive (AR) method are as follows: Figure 5b As shown, the predicted power angle at t = 1.42s is 182.4°, a value obtained at t = 1.39s. The system is determined to be unstable, with a determination time of 0.14s. Figure 5c As shown, compared with the MLE method, the system mode can only be identified at t=1.63s, and the system instability is identified at 0.38s due to MLE>0. The reason for the long time is insufficient speed information, which cannot quickly identify the oscillation mode. Using the method proposed in this invention, t w1 To t w2 Using a data observation window, the rotational speed and power angle are predicted for the next 30 ms at t = 1.35 s. It is predicted that PMLE > 0 at t = 1.61 s, indicating initial power angle instability, which takes 0.10 s. Therefore, the proposed method reduces the time taken by 0.28 s and can accurately identify initial power angle instability.

[0101] Taking fault 2 as an example, the test results of this invention are presented, and the method proposed in this invention is compared with other methods. Figure 6a As shown, using the AR method, the predicted power angle at t = 1.46s is 213.8°, a value obtained at t = 1.43s. The system was incorrectly classified as unstable, with the classification taking 0.18s. Figure 6b As shown, a comparison is made with the MLE method. Based on the rotational speed information, the oscillation pattern is identified, and the MLE can be calculated at t = 2.03s. The first swing peak is reached at t = 3.04s, indicating stable power angle, taking 1.79s. The longer time is due to insufficient power angle information, preventing a rapid arrival at the first swing peak of the MLE. Figure 6c As shown, using the method proposed in this paper, the rotational speed crosses zero at t = 2.16s, so the prediction model switches to II. At t = 2.32s, the PMLE for the next 40ms is predicted. The tangent intercept of the PMLE phase plane trajectory at t = 2.72s, DPMLE = -0.19 < 0, so the power angle is determined to be stable, with a time consumption of 1.07s. Therefore, the proposed method shortens the time consumption by 0.72s and can accurately identify power angle stability.

[0102] As shown in Table 1, extensive tests were conducted on the system to further verify the performance of the PMLE method proposed in this invention in terms of accuracy and time consumption. Specifically, a fault was applied to each near-bus line at t = 1s. The faulted lines were disconnected at 1.20s, 1.25s, 1.30s, 1.35s, and 1.40s, respectively. Figure 7a The time consumed by the MLE method for evaluation Figure 7b The time consumption is evaluated by the PMLE method proposed in this invention. The results of 165 tests show that the relative power angle difference is greater than 180° and the AR method exhibits severe conservatism. In particular, while the AR method significantly shortens the instability identification time, it greatly increases the false positive rate. This is because it only uses the temporal information of the power angle.

[0103] Furthermore, the monitoring data used for prediction even includes information about the duration of the fault, although the power angle motion during these two periods is completely different. The MLE method achieves extremely high discrimination accuracy, but its discrimination process is obviously time-consuming. The proposed PMLE method, compared to the MLE method, only reduces accuracy by 0.61%, but achieves a 55.2% reduction in instability discrimination time and a 45.8% reduction in stability discrimination time. Therefore, the proposed method significantly improves speed while maintaining high accuracy.

[0104] Table 1 Test Results of the New England System

[0105] Evaluation methods δ>180° AR MLE PMLE accuracy 92.7％ 67.9％ 99.4％ 98.8％ Average time to instability (s) 0.08 0.05 0.29 0.13 Stable average time (s) —— 1.58 1.70 0.92 Overall average time elapsed (s) —— 0.72 1.19 0.64

[0106] Since the system stability assessment is affected by the step size of the prediction model, models I and II were further tested with different prediction step sizes to observe the prediction performance. Figure 8 The results show the accuracy and average discrimination time for different prediction step sizes. Predictably, a longer prediction step size results in shorter processing time but lower accuracy. It is evident that when W1 > 30ms and W2 > 40ms, increasing the prediction step size significantly reduces accuracy. Therefore, the chosen prediction step size requires a trade-off between accuracy and processing time.

[0107] To further verify the applicability of this invention to wind power systems, the wind farm in the tested system was put into operation. Buses 3, 4, 16, 17, and 18 were connected to a doubly-fed induction generator (DFIG) wind farm with a rated power of 200MW, and the wind power penetration rate was 16.26%. The fault durations were 0.20s, 0.30s, 0.40s, and 0.50s, respectively, for a total of 132 tests. The test results are shown in Table 2.

[0108] Test results show that the proposed PMLE method can still maintain high accuracy and short processing time in multiple wind farm systems.

[0109] Table 2 Test results of the New England system connecting multiple wind farms

[0110] Clearing time Stable output Output instability Exact number accuracy Average time elapsed (s) 0.20s 33 0 33 100％ 0.88 0.30s 27 6 32 97.0％. 0.75 0.40s 19 14 32 97.0％ 0.52 0.50s 12 21 33 100％ 0.40 total 91 41 130 98.5％ 0.65

[0111] This invention presents an online method for transient power angle stability assessment, enabling rapid online assessment of transient power angle stability after actual fault clearance. The method includes relative speed prediction, relative power angle prediction, PMLE estimation, PMLE phase plane trajectory calculation, and stability assessment. By employing a variable prediction model and self-memory theory, the speed of the MLE method is improved. Test results verify the effectiveness of the proposed method. Notably, this method is lightweight and transferable. Furthermore, it is applicable to systems including wind power. Choosing an appropriate and reliable prediction step size ensures an accuracy of over 98% while significantly reducing processing time. The proposed method can quickly output stability assessment results, providing a basis and signal for emergency measures and saving valuable time, showing broad prospects for large-scale power grid applications.

[0112] According to another aspect of the present invention, a transient power angle stability online evaluation device is provided, comprising:

[0113] The first prediction module is used to predict the relative speed of severely disturbed generator pairs using a prediction model with a variable regression equation after the fault is cleared.

[0114] The second prediction module is used to input the relative power angle and relative speed collected from the severely disturbed generator set into the self-memory prediction model to obtain the predicted relative power angle.

[0115] The acquisition module is used to determine the oscillation mode classification using the relative rotational speed and the relative power angle, and then obtain the corresponding maximum Lyapunov MLE time-domain curve, denoted as PMLE time-domain curve;

[0116] A module is established to calculate the differential value of the PMLE using the discrete data corresponding to the PMLE time-domain curve, so as to establish the phase plane trajectory of the PMLE;

[0117] The evaluation module is used to analyze the PMLE time-domain curve and the PMLE phase plane trajectory to evaluate the stability of the transient power angle.

[0118] According to another aspect of the present invention, a transient power angle stability online evaluation system is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described transient power angle stability online evaluation method.

[0119] According to another aspect of the present invention, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described online evaluation method for transient power angle stability.

[0120] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for online evaluation of transient work angle stability, characterized in that, include: S1: After the fault is cleared, the relative speed of the severely disturbed generator pair is predicted using a prediction model with a variable regression equation. S2: Input the relative power angle and relative speed collected from the severely disturbed generator set pair into the self-memory prediction model to obtain the predicted relative power angle; S3: Use the relative rotational speed and the relative power angle to determine the oscillation mode classification, and then obtain the corresponding maximum Lyapunov MLE time-domain curve, denoted as PMLE time-domain curve; S4: Calculate the differential value of PMLE using the discrete data corresponding to the PMLE time-domain curve to establish the phase plane trajectory of PMLE; S5: Analyze the PMLE time-domain curve and the PMLE phase plane trajectory to evaluate the stability of the transient power angle; In S1: the speed prediction includes: prediction model I based on the exponential function regression model and prediction model II based on the Fourier function prediction model; prediction model I is used after the fault is cleared, and the prediction model II is switched on when the relative speed rises to zero. S5 includes: if the PMLE curve of the severely disturbed generator pair initially decreases, during the first oscillation process, the parameters corresponding to the inflection point and a certain time thereafter of the phase plane trajectory. If the value is greater than 1, the severely disturbed generator set is assessed as having multiple swing instability; if the relative power angle is considered as a cosine function, then... The parameters of the fitted function are obtained by... Calculated t For the current moment, For PMLE value, The parameters of the fitted function are... The angular velocity is the periodic oscillation of the relative work angle. ; The frequency of the periodic oscillation; S5 includes: if the PMLE curve of the severely disturbed generator pair initially decreases, during the first oscillation, the intercept of the phase plane trajectory on the PMLE axis at the inflection point and at a certain time thereafter. D PMLE If the value is less than 0, the corresponding severely disturbed generator set is assessed as stable.

2. The online evaluation method for transient power angle stability as described in claim 1, characterized in that, The prediction model I is represented as follows: ;in, r Let be the order of the regression equation. Let be the regression coefficient of the m-th order, and t be the current time. The relative rotational speed is given.

3. The online evaluation method for transient power angle stability as described in claim 1, characterized in that, The prediction model II is expressed as follows: ; in, r Let be the order of the regression equation. , These are the two regression coefficients of the m-th order. Let t be the frequency of the periodic oscillation, and t be the current time.

4. The online evaluation method for transient power angle stability as described in claim 1, characterized in that, The expression: , A The amplitude of the relative work angle fitted cosine curve is given. t m The time interval between the original trajectory and neighboring trajectories determined for PMLE estimation. and These are the initial points of the original trajectory and the neighboring trajectories, respectively.

5. The online evaluation method for transient power angle stability as described in any one of claims 1-4, characterized in that, Before step S1, the method further includes: After a fault occurs, the rotational speed of each generator set is collected in real time. At the fault clearing time, calculate the absolute value of the difference between the speed of each generator set and the speed of the preset generator set, and use a preset multiple of the maximum absolute value of the difference as the threshold. The absolute value of the difference corresponding to each generator set is compared with the threshold. If it is greater than the threshold, the generator set and the preset generator set are determined to be the severely disturbed generator set pair. The preset generator set is the generator set with the smallest speed change.

6. A transient work angle stability online evaluation device, characterized in that, The method for performing the online evaluation of transient power angle stability according to any one of claims 1-5 includes: The first prediction module is used to predict the relative speed of severely disturbed generator pairs using a prediction model with a variable regression equation after the fault is cleared. The second prediction module is used to input the relative power angle and relative speed collected from the severely disturbed generator set into the self-memory prediction model to obtain the predicted relative power angle. The acquisition module is used to determine the oscillation mode classification using the relative rotational speed and the relative power angle, and then obtain the corresponding maximum Lyapunov MLE time-domain curve, denoted as PMLE time-domain curve; A module is established to calculate the differential value of the PMLE using the discrete data corresponding to the PMLE time-domain curve, so as to establish the phase plane trajectory of the PMLE; The evaluation module is used to analyze the PMLE time-domain curve and the PMLE phase plane trajectory to evaluate the stability of the transient power angle.

7. A transient power angle stability online evaluation system, comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 5.

Images