Power grid overall situation prediction method integrated with electric energy quality event

Abstract

The invention belongs to the field of power grid systems, and particularly relates to a power grid overall situation prediction method integrated with an electric energy quality event. Firstly, a power monitoring system is used for collecting historical data of power quality, and multi-dimensional index data is set to identify related power quality events. And then calculating a cross influence coefficient between different power quality events based on amplitude relevance. Then, an FNN model considering the cross influence coefficient is constructed, that is, Bi-LSTM is adopted as an input layer, adjustment weighting and nonlinear transformation are performed on data according to the cross influence coefficient in a middle layer, and an actual occurrence probability set of each power quality event is output in an output layer; and finally, constructing a power grid global situation map, and performing thermodynamic diagram analysis to predict the development trend of the power grid global situation. According to the scheme, multi-dimensional power quality indexes are considered, the global situation of the power grid can be accurately evaluated, and powerful support is provided for operation and management of the power grid.

Description

Technical Field

[0001] The present invention belongs to the field of power grid systems, and in particular relates to a method for predicting the overall situation of a power grid incorporating power quality events. Background Art

[0002] In the operation and management of power systems, power quality issues have always been a key factor affecting grid stability and power supply reliability. With the rapid development of industrialization and informatization, the power quality issues faced by power systems have become more complex and changeable. Traditional power quality analysis methods often focus on single indicators, such as voltage stability, frequency deviation, etc. These methods can provide a limited perspective to observe and solve specific power quality problems. However, the actual situation of power quality is the result of the combined effect of multiple factors, including but not limited to the operating conditions of the power grid, load changes, equipment aging, and natural factors.

[0003] The limitation of single indicator analysis methods is that they cannot fully capture the multi-dimensional characteristics of the grid operation status, nor can they effectively predict power quality events that may be caused by the interaction of multiple factors. For example, voltage sag may be associated with harmonic pollution in the grid, while three-phase imbalance may affect the frequency stability of the grid. In addition, with the large-scale grid connection of renewable energy, the operation mode and power quality characteristics of the grid are also changing, which requires more advanced methods to adapt to these new challenges. Summary of the invention

[0004] In view of the shortcomings of the existing technology, the present invention proposes a method for predicting the overall situation of the power grid that incorporates power quality events. Historical data is collected through the power monitoring system, multi-dimensional power quality indicators are set, the cross-influence coefficient of power quality events is calculated, and the Bi-LSTM and FNN models are used for prediction. A situation diagram is constructed by combining the topological structure and equipment parameters to predict the development trend, so as to comprehensively evaluate the power quality and situation of the power grid.

[0005] The technical solution of the present invention is as follows:

[0006] A method for predicting the overall state of a power grid incorporating power quality events, comprising:

[0007] Use the power monitoring system to collect historical data on power quality, analyze the historical data, set multi-dimensional indicator data on power quality, analyze the historical data based on the multi-dimensional indicator data, and identify power quality events related to power quality in power grid operation;

[0008] Calculate the cross-influence coefficient between different power quality events based on amplitude correlation;

[0009] Construct an FNN model that takes the cross-influence coefficient into consideration. The construction method uses Bi-LSTM as the input layer to input historical data. In the middle layer, the historical data is adjusted and weighted according to the cross-influence coefficient, and nonlinear transformation is performed. In the output layer, the actual occurrence probability set of each power quality event is output;

[0010] According to the actual occurrence probability set output by the FNN model, the changing trend of the probability of power quality events is analyzed, and the global situation map of the power grid is constructed in combination with the topological structure and equipment parameters of the power grid. The global situation map of the power grid is analyzed by thermal diagram to predict the development trend of the global situation of the power grid.

[0011] Furthermore, the multi-dimensional indicator data of the power quality are set to include: voltage deviation amplitude and duration, frequency deviation and duration, total harmonic distortion and three-phase imbalance;

[0012] Among them, the voltage deviation amplitude is: Among them, V min is the minimum voltage value, V nom is the rated voltage;

[0013] Wherein, frequency deviation: Δf=|f m -f n |, where f m is the measurement frequency, f n is the nominal frequency;

[0014] Among them, the total harmonic distortion is: Among them, V n is the voltage effective value of the nth harmonic, and V1 is the voltage effective value of the fundamental wave;

[0015] Among them, the three-phase unbalance degree is: Among them, V a ,V b ,V c It is the effective value of the three-phase voltage in a three-phase power system.

[0016] Further, the identifying power quality event includes:

[0017] Voltage sag: Set the voltage sag to 90% of the rated voltage and last for more than 10 milliseconds;

[0018] Voltage swell: Set the voltage swell to 110% of the rated voltage and last for more than 10 milliseconds;

[0019] Voltage interruption: The set voltage RMS value drops to 0 and lasts for more than 0.5 seconds;

[0020] Voltage fluctuation: Set the voltage fluctuation amplitude to ±5% of the rated voltage, and the duration threshold to 1 minute;

[0021] Harmonic pollution: total harmonic loss does not exceed 5%;

[0022] Three-phase unbalance: Set the three-phase unbalance to exceed 5%.

[0023] Furthermore, the formula for calculating the cross-influence coefficient between different power quality events based on amplitude correlation is:

[0024]

[0025] Where, suppose the change sequence of power quality event A is A={a1,a2,a3,...,a n}, the change sequence of power quality event B is B={b1,b2,b3,...,b n},a1,a2,a3,...,a n and b1,b2,b3,...,b n are the change amplitudes of power quality event A and power quality event B respectively; r AB represents the correlation coefficient, with a value range between -1 and 1; n represents the number of change sequences, a i and b i They respectively represent the change amplitude values ​​of power quality event A and power quality event B at the i-th data point; and They represent the average values ​​of the change amplitudes of power quality event A and power quality event B, that is, The weight of power quality event A is ω A , the weight of power quality event B is ω B .

[0026] Furthermore, the cross-influence coefficient performs weighted distribution on the input data processed by Bi-LSTM, and the weighted distribution formula is as follows:

[0027] x′ A =x A ·(1+C AB );

[0028] x′ B =x B ·(1+C AB );

[0029] Among them, x′ A and x′ B represents the weighted distribution data of the input power quality event A and the power quality event B; C AB Represents the cross-influence coefficient between power quality event A and power quality event B; x A and x BRepresents the data of power quality event A and power quality event B after being processed by Bi-LSTM.

[0030] Furthermore, the output after the nonlinear transformation is:

[0031] x′ A ′=ω′ A x′ A +b A ;

[0032] x′ B ′=ω′ B x′ B +b B ;

[0033] Among them, x′ A ′ and x′ B ′ represents the data of power quality event A and power quality event B after nonlinear transformation; ω′ A and ω′ B The connection weights between the intermediate layer neurons and the input layer neurons for power quality event A and power quality event B; b A and b B Represents the bias term of the intermediate layer neurons of power quality event A and power quality event B.

[0034] Furthermore, the actual occurrence probability set calculation formula output by the FNN model is:

[0035]

[0036] Among them, x i is the power quality event data after being processed by the previous layers; i represents different power quality events; e represents a natural constant; j represents traversing all power quality events.

[0037] The beneficial effect of the present invention is that by setting multi-dimensional power quality indicators, including voltage deviation amplitude and duration, frequency deviation and duration, total harmonic distortion and three-phase imbalance, the power quality status of the power grid can be more comprehensively evaluated.

[0038] Calculating the cross-influence coefficients between different power quality events based on amplitude correlation can more accurately reflect the relationship between power quality events and provide more valuable information for subsequent analysis and prediction.

[0039] Using Bi-LSTM as the input layer can better capture the time series characteristics and long-term dependencies in historical data, and improve the accuracy of the model in predicting power quality events. Constructing an FNN model that takes into account the cross-influence coefficient can adjust the weighting and nonlinear transformation of historical data according to the cross-influence coefficient in the middle layer, further improving the performance of the model.

[0040] Combining the topological structure and equipment parameters of the power grid can provide a more comprehensive understanding of the operating status of the power grid, analyze the impact of the topological structure and equipment parameters on power quality events, and provide more targeted decision-making support for the operation and management of the power grid.

[0041] By analyzing the changing trend of the probability of power quality events, constructing a global power grid situation map, and conducting thermal map analysis, it is possible to predict the development trend of the global power grid situation, take measures in advance to deal with possible problems, and ensure the stable operation of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 It is an overall flow chart of a method for predicting the overall situation of a power grid incorporating power quality events;

[0043] Figure 2 Provides specific steps for data collection, multi-dimensional indicator setting and power quality events;

[0044] Figure 3 Steps for constructing the FNN model considering the cross-influence coefficient;

[0045] Figure 4 These are the specific implementation steps of S400. DETAILED DESCRIPTION

[0046] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0047] See also Figure 1 , which shows a method for predicting the overall situation of a power grid incorporating power quality events provided by an embodiment of the present invention, comprising:

[0048] S100: Utilize the power monitoring system to collect historical data on power quality, analyze the historical data, set multi-dimensional indicator data on power quality, analyze the historical data based on the multi-dimensional indicator data, and identify power quality events related to power quality in power grid operation.

[0049] S200: Calculate the cross-influence coefficients between different power quality events based on the amplitude correlation.

[0050] S300: Construct an FNN model that takes the cross-influence coefficient into consideration. The construction method uses Bi-LSTM as the input layer to input historical data. In the middle layer, the historical data is adjusted and weighted according to the cross-influence coefficient, and a nonlinear transformation is performed. In the output layer, the actual occurrence probability set of each power quality event is output.

[0051] S400: Analyze the changing trend of the probability of power quality events based on the actual occurrence probability set output by the FNN model, and build a global situation map of the power grid in combination with the topological structure and equipment parameters of the power grid; perform thermal map analysis on the global situation map of the power grid to predict the development trend of the global situation of the power grid.

[0052] S100 involves using a power monitoring system to collect historical data on power quality, analyzing the historical data to identify events related to power quality in power grid operation, and calculating multi-dimensional indicator data of power quality through the data analysis module of the power monitoring system. Specifically, this process involves data collection, storage, preprocessing, and the use of specific algorithms and calculation formulas to identify power quality events and calculate multi-dimensional indicator data such as voltage deviation, frequency deviation, and harmonic content. In this step, the power monitoring system is a system for real-time monitoring of the operating status of the power grid, which can collect and process various power parameters and signals in the power grid. The power monitoring system generally includes sensors, data acquisition equipment, communication networks, and data analysis modules. The data analysis module in the power monitoring system is responsible for processing and analyzing the collected data.

[0053] In some possible implementations, such as Figure 2 As shown, the specific contents of S100 include:

[0054] S110: Collect, store and preprocess data using the power monitoring system. The sensors and monitoring equipment in the power monitoring system collect electrical parameter data such as voltage, current, frequency, etc. at each node in the power grid in real time. The collected data is stored in the database for subsequent data analysis and processing. The stored data is cleaned and screened to remove abnormal data and noise to ensure the accuracy and reliability of the data.

[0055] S120: According to the characteristics and requirements of the power grid, multi-dimensional indicator data of power quality are set.

[0056] The multi-dimensional indicator data of power quality are set, including: voltage deviation amplitude and duration, frequency deviation and duration, total harmonic distortion and three-phase imbalance;

[0057] Among them, the voltage deviation amplitude is: Among them, V min is the minimum voltage value, V nom is the rated voltage.

[0058] Wherein, frequency deviation: Δf=|f m -f n |, where f m is the measurement frequency, f n is the nominal frequency.

[0059] Among them, the total harmonic distortion is: Among them, V n is the effective value of the voltage of the nth harmonic, and V1 is the effective value of the voltage of the fundamental wave.

[0060] Among them, the three-phase unbalance degree is: Among them, V a ,V b ,V c It is the effective value of the three-phase voltage in a three-phase power system.

[0061] S130: Analyze historical data based on multi-dimensional indicator data to identify power quality events related to power quality in power grid operation.

[0062] The identified power quality events are as follows:

[0063] Voltage sag: Set the voltage sag to 90% of the rated voltage. If the effective voltage value is lower than this value and lasts for more than 10 milliseconds, it is considered that a voltage sag event has occurred.

[0064] Voltage swell: Set the voltage swell to 110% of the rated voltage. When the voltage effective value exceeds this value and lasts for more than 10 milliseconds, a voltage swell event is considered to have occurred.

[0065] Voltage interruption: The voltage RMS value is set to drop to 0 and lasts for more than 0.5 seconds. If the voltage fluctuation meets these conditions, it can be determined as a voltage interruption event.

[0066] Voltage fluctuation: Set the voltage fluctuation amplitude to ±5% of the rated voltage and the duration threshold to 1 minute. If the voltage fluctuation meets these conditions, it can be determined as a voltage fluctuation event.

[0067] Harmonic pollution: The total harmonic distortion should not normally exceed 5%. If the actual total harmonic distortion exceeds 5%, it can be considered that a harmonic pollution event has occurred.

[0068] Three-phase unbalance: When the three-phase unbalance exceeds 5%, it is considered that a three-phase unbalance event has occurred.

[0069] S200 involves calculating the cross-influence coefficients between different power quality events based on amplitude correlation. For power quality events with amplitude characteristics, it is analyzed whether the amplitudes of other power quality events also show corresponding change trends when one power quality event occurs.

[0070] The specific calculation method is as follows:

[0071] For the associated power quality events, select the historical data sample points where they occur simultaneously within a period of time, and assume that the change sequence of power quality event A is A = {a1, a2, a3, ..., a n}, the change sequence of power quality event B is B={b1,b2,b3,...,b n}, where a1, a2, a3, ..., a n and b1,b2,b3,...,b n They are the change amplitude values ​​of power quality event A and power quality event B respectively.

[0072] Calculate the correlation coefficient between power quality event A and power quality event B. The calculation formula is:

[0073]

[0074] Among them, r AB represents the correlation coefficient, with a value range between -1 and 1; n represents the number of change sequences, a i and b i They respectively represent the change amplitude values ​​of power quality event A and power quality event B at the i-th data point; and They represent the average values ​​of the change amplitudes of power quality event A and power quality event B, that is,

[0075] Construct a correction function and set the weight of power quality event A to be ω A , the weight of power quality event B is ω B , then the correction function

[0076] The calculation formula of cross-influence coefficient is:

[0077]

[0078] S300 mainly builds an FNN model that considers the cross-influence coefficient. First, the FNN network is improved, and Bi-LSTM is used as the input layer to input historical data to capture the time series characteristics. Then, the data is adjusted and weighted in the middle layer according to the cross-influence coefficients of different power quality events, and nonlinear transformation is performed. Finally, the actual occurrence probability set of each power quality event is output in the output layer, and the Softmax function is used for probability distribution conversion. Through these steps, the model can comprehensively consider the cross-influence and more accurately predict the probability of power quality events.

[0079] In some possible embodiments, such as Figure 3 As shown, S300 specifically includes:

[0080] S301: Use Bi-LSTM as the input layer to input historical data. Bi-LSTM can consider information in both the forward and backward directions at the same time and better capture the temporal characteristics in historical data.

[0081] A bidirectional long short-term memory network (Bi-LSTM) is used to process historical data input. Bi-LSTM can simultaneously consider the information in both the forward and backward directions of historical data, so as to better capture the temporal features and long-term dependencies in the data. It processes data in two directions: forward propagation and backward propagation. In forward propagation, the forget gate determines which information is discarded from the cell state, the input gate determines which new information will be stored in the cell state, the candidate cell state represents the possible new cell state, the cell state is updated according to the results of the forget gate and the input gate, and the output gate determines the output hidden state at the current moment. In backward propagation, similar calculations are performed, but in the opposite direction, processing data from back to front.

[0082] After the historical data is input into Bi-LSTM, the output is a hidden state sequence that takes into account the previous and next time series information through the forward and backward propagation of Bi-LSTM. Using Bi-LSTM as the input layer can make full use of the time series information of historical data, provide more effective input for the cross-influence coefficient allocation and nonlinear transformation of the subsequent intermediate layer, and thus improve the accuracy of the model in predicting power quality events.

[0083] S302: Perform weighted distribution on the input data processed by the Bi-LSTM according to the cross-influence coefficients between different power quality events calculated in S200.

[0084] The weighted allocation formula is as follows:

[0085] x′ A =x A ·(1+C AB );

[0086] x′ B =x B ·(1+C AB );

[0087] Among them, C AB Represents the cross-influence coefficient x between power quality event A and power quality event B A and x B Represents the data of power quality event A and power quality event B after being processed by Bi-LSTM.

[0088] S303: Perform nonlinear transformation in the middle layer to enhance the expressive power of the model.

[0089] In the middle layer, for the input power quality event A and power quality event B, the weighted distribution of x′ A and x′ B , the output after nonlinear transformation is:

[0090] x′ A ′=ω′ A x′ A +b A ;

[0091] x′ B ′=ω′ B x′ B +b B ;

[0092] Among them, x′ A ′ and x′ B ′ represents the data of power quality event A and power quality event B after nonlinear transformation; ω′ A and ω′ B The connection weights between the intermediate layer neurons and the input layer neurons for power quality event A and power quality event B; b A and b B Represents the bias term of the intermediate layer neurons of power quality event A and power quality event B.

[0093] S304: The output layer outputs the actual occurrence probability set of each power quality event. Through the processing and learning of the previous layers, the model can predict the probability of each event based on the input historical data.

[0094] The formula for calculating the probability of occurrence is:

[0095]

[0096] Among them, x iis the power quality event data after being processed by the previous layers; i represents different power quality events; e represents a natural constant; j represents traversing all power quality events.

[0097] S400 combines the grid topology and equipment parameters with the changing trends of the probability of power quality events to analyze the relationship between them. It then constructs a global grid situation map based on the analysis results and intuitively displays the power quality conditions in different areas through thermal map analysis. Finally, based on the results of the thermal map analysis, it predicts the development trend of the global grid situation to provide support for the operation and management of the grid.

[0098] In some possible embodiments, Figure 4 As shown, S400 specifically includes:

[0099] S401: Obtain the topological structure information of the power grid, including the connection relationship between each node and line, and the parameters of the equipment, such as transformer capacity, line impedance, etc. Combine this information with the changing trend of the probability of power quality events to more comprehensively understand the operating status of the power grid.

[0100] Specifically, the changing trend of the probability of power quality events is associated with the topological structure and equipment parameters of the power grid. The corresponding relationship between power quality data and each node, line and equipment in the power grid is established. Analyze how the topological structure and equipment parameters affect the occurrence and propagation of power quality events. For example, insufficient transformer capacity may cause voltage sag, and excessive line impedance may cause voltage fluctuations. Evaluate the power quality characteristics of different regions based on the topological structure and equipment parameters. Determine which areas may be more susceptible to specific power quality events and the propagation paths of these events in the power grid. Based on the above analysis, provide decision support for the operation and management of the power grid.

[0101] S402: Based on the analysis results of the changing trends of the power grid topology, equipment parameters and the probability of power quality events, a power grid global situation map is constructed; the situation map can graphically display the power quality status, potential risks and weak links of each area in the power grid. It can be presented in the form of maps or charts.

[0102] S403: Perform a heat map analysis on the global situation map of the power grid, and use color coding to indicate the power quality status of different areas. The heat map can help identify hot spots and abnormal situations, so as to more intuitively understand the operation status of the power grid, and then predict the development trend of the global situation of the power grid.

[0103] The basic principles of the present application are described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, effects, etc. mentioned in the present application are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. are required by each embodiment of the present application. In addition, the specific details disclosed above are only for the purpose of illustration and ease of understanding, not for limitation, and the above details do not limit the present application to being implemented by adopting the above specific details.

[0104] The block diagrams of the devices, apparatuses, equipment, and systems involved in this application are only illustrative examples and are not intended to require or imply that they must be connected, arranged, and configured in the manner shown in the block diagram. As will be appreciated by those skilled in the art, these devices, apparatuses, equipment, and systems can be connected, arranged, and configured in any manner. Words such as "including", "comprising", "having", etc. are open words, referring to "including but not limited to", and can be used interchangeably with them. The words "or" and "and" used here refer to the words "and / or" and can be used interchangeably with them, unless the context clearly indicates otherwise. The words "such as" used here refer to the phrase "such as but not limited to", and can be used interchangeably with them.

[0105] It should also be noted that in the apparatus, device and method of the present application, each component or each step can be decomposed and / or recombined. Such decomposition and / or recombination should be regarded as equivalent solutions of the present application.

[0106] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of the present application. Therefore, the present application is not intended to be limited to the aspects shown herein, but rather to the widest scope consistent with the principles and novel features of the present invention.

[0107] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for predicting the overall situation of a power grid incorporating power quality events, characterized in that: include: Use the power monitoring system to collect historical data on power quality, analyze the historical data, set multi-dimensional indicator data on power quality, analyze the historical data based on the multi-dimensional indicator data, and identify power quality events related to power quality in power grid operation; Calculate the cross-influence coefficient between different power quality events based on amplitude correlation; Construct an FNN model that takes the cross-influence coefficient into consideration. The construction method uses Bi-LSTM as the input layer to input historical data. In the middle layer, the historical data is adjusted and weighted according to the cross-influence coefficient, and nonlinear transformation is performed. In the output layer, the actual occurrence probability set of each power quality event is output; According to the actual occurrence probability set output by the FNN model, the changing trend of the probability of power quality events is analyzed, and the global situation map of the power grid is constructed in combination with the topological structure and equipment parameters of the power grid. The global situation map of the power grid is analyzed by thermal diagram to predict the development trend of the global situation of the power grid.

2. A method for predicting the overall situation of a power grid incorporating power quality events according to claim 1, characterized in that: The multi-dimensional indicator data of the power quality are set to include: voltage deviation amplitude and duration, frequency deviation and duration, total harmonic distortion and three-phase imbalance; Among them, the voltage deviation amplitude is: Where λ is the voltage deviation amplitude, V min is the minimum voltage value, V nom is the rated voltage; Where, frequency deviation: Δf=|f m -f n |, where Δf is the frequency deviation, f m is the measurement frequency, f n is the nominal frequency; Among them, the total harmonic distortion is: Where THD is the total harmonic distortion, V n is the voltage effective value of the nth harmonic, and V1 is the voltage effective value of the fundamental wave; Among them, the three-phase unbalance degree is: Among them, Ub is the three-phase unbalance, V a ,V b ,V c It is the effective value of the three-phase voltage in a three-phase power system.

3. A method for predicting the overall situation of a power grid incorporating power quality events according to claim 2, characterized in that: Identifying the power quality event includes: Voltage sag: Set the voltage sag to 90% of the rated voltage and last for more than 10 milliseconds; Voltage swell: Set the voltage swell to 110% of the rated voltage and last for more than 10 milliseconds; Voltage interruption: The set voltage RMS value drops to 0 and lasts for more than 0.5 seconds; Voltage fluctuation: Set the voltage fluctuation amplitude to ±5% of the rated voltage, and the duration threshold to 1 minute; Harmonic pollution: total harmonic loss does not exceed 5%; Three-phase unbalance: Set the three-phase unbalance to exceed 5%.

4. A method for predicting the overall situation of a power grid incorporating power quality events according to claim 1, characterized in that: The formula for calculating the cross-influence coefficient between different power quality events based on amplitude correlation is: Among them, C AB is the cross-influence coefficient; suppose the change sequence of power quality event A is A={a1,a2,a3,...,a n }, the change sequence of power quality event B is B={b1,b2,b3,...,b n },a1,a2,a3,...,a n and b1,b2,b3,...,b n are the change amplitudes of power quality event A and power quality event B respectively; r AB represents the correlation coefficient, with a value range between -1 and 1; n represents the number of change sequences, a i and b i They respectively represent the change amplitude values ​​of power quality event A and power quality event B at the i-th data point; and They represent the average values ​​of the change amplitudes of power quality event A and power quality event B, that is, The weight of power quality event A is ω A , the weight of power quality event B is ω B .

5. A method for predicting the overall situation of a power grid incorporating power quality events according to claim 4, characterized in that: The cross-influence coefficient performs weighted distribution on the input data after Bi-LSTM processing, and the weighted distribution formula is as follows: x′ A =x A ·(1+C AB ); x′ B =x B ·(1+C AB ); Among them, x′ A and x′ B represents the weighted distribution data of the input power quality event A and the power quality event B; C AB Represents the cross-influence coefficient between power quality event A and power quality event B; x A and x B Represents the data of power quality event A and power quality event B after being processed by Bi-LSTM.

6. A method for predicting the overall situation of a power grid incorporating power quality events according to claim 5, characterized in that: The output after the nonlinear transformation is: x′ A ′=ω′ A x′ A +b A ; x′ B ′=ω′ B x′ B +b B ; Among them, x′ A ′ and x′ B ′ represents the data of power quality event A and power quality event B after nonlinear transformation; ω′ A and ω′ B The connection weights between the intermediate layer neurons and the input layer neurons for power quality event A and power quality event B; b A and b B Represents the bias term of the intermediate layer neurons of power quality event A and power quality event B.

7. A method for predicting the overall situation of a power grid incorporating power quality events according to claim 6, characterized in that: The actual occurrence probability set calculation formula output by the FNN model is: Among them, p i is the actual occurrence probability set output by the FNN model, x i is the power quality event data after being processed by the previous layers; i represents different power quality events; e represents a natural constant; j represents traversing all power quality events.

Images