Risk analysis method and device in power grid operation process

Abstract

The invention discloses a risk analysis method and device in a power grid operation process. The method comprises the following steps: preprocessing current power grid state data of a target power grid in an operation process to obtain preprocessed current power grid state data; performing feature extraction on the preprocessed current power grid state data to obtain feature data of a target power grid; processing the feature data by using a power grid state data prediction model to obtain future power grid state data of the target power grid in a future predetermined time period; determining a predetermined index of the target power grid by using the future power grid state data; determining a risk index of the target power grid according to the index value corresponding to the predetermined index; and determining the risk level of the target power grid according to the risk index. According to the method and the device, the technical problems that the risk analysis mode for the operation process of the power grid in the related technology can only analyze historical data and cannot respond to the change of the operation state of the power grid and the dynamic change of the risk in time, and the reliability is relatively low are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid operation risk analysis, and in particular to a method and device for risk analysis during power grid operation. Background Art

[0002] The current power system is expanding, the grid structure is becoming increasingly complex, and the large-scale access of new energy sources such as wind power and photovoltaics has also brought new challenges to grid operation. Traditional grid operation risk analysis methods are mainly based on historical data and empirical rules, which have the following problems.

[0003] For example, there is a lack of comprehensive consideration of multiple factors. The risk of power grid operation is affected by multiple factors, including power grid status, new energy forecasts, load changes, etc. Traditional methods find it difficult to comprehensively consider these factors, resulting in incomplete analysis results. Lack of pertinence and accuracy. Traditional methods find it difficult to fully consider the impact of factors such as power grid status and new energy forecasts on power grid operation risks, resulting in a lack of pertinence and accuracy in the analysis results, and an inability to respond to risk changes in a timely manner. Traditional methods can usually only analyze historical data and cannot respond to changes in power grid operation status and dynamic changes in risks in a timely manner. Therefore, it is necessary to provide an efficient and accurate power grid operation risk analysis method to meet the new challenges of power grid operation management and improve the safety and stability of power grid operation.

[0004] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0005] An embodiment of the present invention provides a method and device for risk analysis during the operation of a power grid, so as to at least solve the technical problem that the method used in the related art for risk analysis of the operation of a power grid can usually only analyze historical data, cannot respond in time to changes in the operation status of the power grid and dynamic changes in risks, and has low reliability.

[0006] According to one aspect of an embodiment of the present invention, a risk analysis method in a power grid operation process is provided, comprising: preprocessing current power grid state data of a target power grid during operation to obtain the preprocessed current power grid state data; extracting features of the preprocessed current power grid state data to obtain feature data of the target power grid; processing the feature data using a power grid state data prediction model to obtain future power grid state data of the target power grid in a predetermined future period, wherein the power grid state data prediction model is obtained by machine learning training using multiple sets of training data, each of the multiple sets of training data comprising: sample feature data and sample future power grid state data corresponding to the sample feature data; determining predetermined indicators of the target power grid using the future power grid state data, wherein the predetermined indicators at least comprise: an expected load loss indicator, a power grid state indicator, and an over-limit indicator, wherein the expected load loss indicator represents the proportion of the load required to be removed before and after a fault occurs in the target power grid to the total load, and the over-limit indicator represents the operating state of the target power grid at a predetermined time; determining a risk index of the target power grid according to the indicator value corresponding to the predetermined indicator; and determining the risk level of the target power grid according to the risk index.

[0007] Optionally, before preprocessing the current grid status data of the target power grid during operation to obtain the preprocessed current grid status data, the risk analysis method during the power grid operation process also includes: when it is determined that the target power grid enters the operating state, triggering a first sensor on each grid circuit in the target power grid and a second sensor arranged at a predetermined position of each electrical equipment in the target power grid to start, so as to collect the current grid status data of the target power grid; using an Internet of Things device to obtain the current grid status data from the first sensor and the second sensor, wherein the current grid status data includes: operating parameters, grid topology and equipment status of the target power grid, new energy power generation data in the target power grid and historical fault data of the target power grid.

[0008] Optionally, the current power grid status data of the target power grid during operation is preprocessed to obtain the preprocessed current power grid status data, including at least one of the following: interpolating missing values ​​in the current power grid status data to obtain the interpolated current power grid status data; filtering noise data in the current power grid status data to obtain the filtered current power grid status data; performing dimensionality reduction processing on the current power grid status data to obtain the dimensionality reduction processing of the current power grid status data.

[0009] Optionally, interpolation processing is performed on the missing values ​​in the current power grid status data to obtain the current power grid status data after interpolation processing, including: determining the interpolation step size of the interpolation processing according to the total amount of data of the current power grid status data; determining the curve coefficient of the spline curve used for the interpolation processing; determining the interpolation data of the missing values ​​according to the interpolation step size, the curve coefficient and the current power grid status data; and inserting the interpolation data into the current power grid status data according to the interpolation step size to obtain the current power grid status data after interpolation processing.

[0010] Optionally, the noise data in the current power grid state data is filtered to obtain the filtered current power grid state data, including: decomposing the state signal corresponding to the current power grid state data by using a variational mode decomposition method to obtain the modal function corresponding to the current power grid state data; denoising the modal function by using a modal filter to obtain the denoised state signal; and obtaining the filtered current power grid state data according to the denoised state signal.

[0011] Optionally, the current power grid state data is subjected to dimensionality reduction processing to obtain the current power grid state data after dimensionality reduction processing, including: inputting the current power grid state data into a target encoder, so as to use the target encoder to perform dimensionality reduction processing on the current power grid state data, so as to obtain the current power grid state data after dimensionality reduction processing, wherein the target encoder is an encoder obtained by the following training of the original encoder: inputting sample power grid state data into the original encoder, so as to use the original encoder to process the sample power grid state data, so as to obtain processed sample coded data; inputting the processed sample coded data into a decoder, so as to use the decoder to process the processed sample coded data, so as to obtain decoded sample coded data; using a loss function to calculate the data difference between the processed sample coded data and the decoded sample coded data; when it is determined based on the data difference that the training convergence condition is met, the target encoder is obtained.

[0012] Optionally, the characteristic data is processed using a power grid state data prediction model to obtain future power grid state data of the target power grid within a predetermined time period in the future, including: using a convolutional neural network to extract features of the current power grid state data to convert the time series data corresponding to the current power grid state data into a current power grid state data sequence with spatial characteristics; obtaining time dependency data in the current power grid state data based on the current power grid state data sequence; formatting the time dependency data to obtain the time dependency data identified by an integrated learning algorithm; and using the integrated learning algorithm to predict the future power grid state data based on the time dependency data and the current power grid state data sequence with spatial characteristics.

[0013] Optionally, determining the risk index of the target power grid according to the index value corresponding to the predetermined index includes: determining the index value one corresponding to the over-limit index by a first formula, wherein the first formula is: M0 represents the index value 1, n represents the number of buses in the target power grid, k represents the kth bus in the target power grid, Indicates the voltage over-limit situation of the kth bus; determine the index value 2 corresponding to the expected load loss index by the second formula, wherein the second formula is: M Ll =∑ c∈M p c *L c , M Ll Indicates the index value 2, L c represents the load lost by the target power grid during operation, p c represents the risk operation probability of the target power grid in the cth state; the indicator value three corresponding to the power grid state indicator is determined by the third formula, wherein the third formula is: M Gs =∑ c∈Gsi p c , M Gs represents the indicator value three, Gsi represents the set of the target power grid being in a predetermined operating state i; the risk index is determined according to the indicator value one, the indicator value two and the indicator value three.

[0014] Optionally, determining the risk index according to the indicator value one, the indicator value two and the indicator value three includes: respectively obtaining the weight values ​​of the indicator value one, the indicator value two and the indicator value three; determining the risk index according to the indicator value one, the indicator value two and the indicator value three and their respective corresponding weight values.

[0015] According to another aspect of an embodiment of the present invention, a risk analysis device in a power grid operation process is also provided, comprising: a preprocessing module, used to preprocess current power grid state data of a target power grid during operation to obtain the preprocessed current power grid state data; a feature extraction module, used to extract features of the preprocessed current power grid state data to obtain feature data of the target power grid; a processing module, used to process the feature data using a power grid state data prediction model to obtain future power grid state data of the target power grid within a predetermined period of time in the future, wherein the power grid state data prediction model is obtained by machine learning training using multiple sets of training data, and the multiple sets of training data Each group of the data includes: sample characteristic data and sample future power grid state data corresponding to the sample characteristic data; a first determination module is used to determine the predetermined indicators of the target power grid using the future power grid state data, wherein the predetermined indicators at least include: an expected load loss indicator, a power grid state indicator and an over-limit indicator, the expected load loss indicator represents the proportion of the load required to be removed before and after the target power grid fails to the total load, and the over-limit indicator represents the operating status of the target power grid at a predetermined time; a second determination module is used to determine the risk index of the target power grid according to the indicator value corresponding to the predetermined indicator; a third determination module is used to determine the risk level of the target power grid according to the risk index.

[0016] Optionally, the risk analysis device during the operation of the power grid also includes: a collection module, which is used to preprocess the current power grid status data of the target power grid during the operation process, and before obtaining the preprocessed current power grid status data, when it is determined that the target power grid enters the operating state, trigger the first sensor on each power grid circuit in the target power grid and the second sensor set at a predetermined position of each electrical equipment in the target power grid to start, so as to collect the current power grid status data of the target power grid; an acquisition module, which is used to use the Internet of Things device to obtain the current power grid status data from the first sensor and the second sensor, wherein the current power grid status data includes: the operating parameters, power grid topology and equipment status of the target power grid, the new energy power generation data in the target power grid and the historical fault data of the target power grid.

[0017] Optionally, the preprocessing module includes at least one of the following: an interpolation unit, used to interpolate missing values ​​in the current power grid status data to obtain the current power grid status data after interpolation; a filtering unit, used to filter noise data in the current power grid status data to obtain the current power grid status data after filtering; a dimensionality reduction unit, used to perform dimensionality reduction on the current power grid status data to obtain the current power grid status data after dimensionality reduction.

[0018] Optionally, the interpolation unit includes: a first determination subunit, used to determine the interpolation step size of the interpolation processing according to the total amount of data of the current power grid status data; a second determination subunit, used to determine the curve coefficient of the spline curve used for the interpolation processing; a third determination subunit, used to determine the interpolation data of the missing value according to the interpolation step size, the curve coefficient and the current power grid status data; an insertion subunit, used to insert the interpolation data into the current power grid status data according to the interpolation step size to obtain the current power grid status data after interpolation processing.

[0019] Optionally, the filtering unit includes: a decomposition subunit, used to decompose the state signal corresponding to the current power grid state data using a variational mode decomposition method to obtain a mode function corresponding to the current power grid state data; a denoising subunit, used to denoise the mode function using a modal filter to obtain the denoised state signal; and a first acquisition subunit, used to obtain the filtered current power grid state data based on the denoised state signal.

[0020] Optionally, the dimension reduction unit includes: a dimension reduction sub-unit, used to input the current power grid state data into a target encoder, so as to use the target encoder to perform dimension reduction processing on the current power grid state data, so as to obtain the current power grid state data after dimension reduction processing, wherein the target encoder is an encoder obtained by the following training of the original encoder: inputting sample power grid state data into the original encoder, so as to use the original encoder to process the sample power grid state data, so as to obtain processed sample coded data; inputting the processed sample coded data into a decoder, so as to use the decoder to process the processed sample coded data, so as to obtain decoded sample coded data; using a loss function to calculate the data difference between the processed sample coded data and the decoded sample coded data; when it is determined based on the data difference that the training convergence condition is met, the target encoder is obtained.

[0021] Optionally, the processing module includes: a feature extraction unit, used to extract features from the current power grid status data using a convolutional neural network, so as to convert the time series data corresponding to the current power grid status data into a current power grid status data sequence with spatial characteristics; an acquisition unit, used to acquire time dependency data in the current power grid status data based on the current power grid status data sequence; a format conversion unit, used to perform format conversion on the time dependency data, so as to obtain the time dependency data identified by an integrated learning algorithm; and a prediction unit, used to predict the future power grid status data based on the time dependency data and the current power grid status data sequence with spatial characteristics using the integrated learning algorithm.

[0022] Optionally, the second determination module includes: a first determination unit, configured to determine an index value 1 corresponding to the over-limit index by using a first formula, wherein the first formula is: M0 represents the index value 1, n represents the number of buses in the target power grid, k represents the kth bus in the target power grid, Indicates the voltage over-limit situation of the kth bus; a second determination unit, used to determine the index value 2 corresponding to the expected load loss index through a second formula, wherein the second formula is: M Ll =∑ c∈M p c *L c , M Ll Indicates the index value 2, L c represents the load lost by the target power grid during operation, p c represents the risk operation probability of the target power grid in the cth state; a third determination unit, used to determine the indicator value three corresponding to the power grid state indicator through a third formula, wherein the third formula is: M Gs =∑ c∈Gsi p c , M Gs represents the indicator value three, Gsi represents the set of the target power grid being in a predetermined operating state i; a fourth determination unit is used to determine the risk index according to the indicator value one, the indicator value two and the indicator value three.

[0023] Optionally, the fourth determination unit includes: a second acquisition subunit, used to respectively obtain the weight values ​​of the indicator value one, the indicator value two and the indicator value three; a fourth determination subunit, used to determine the risk index based on the indicator value one, the indicator value two and the indicator value three and their respective corresponding weight values.

[0024] According to another aspect of an embodiment of the present invention, a computer program product is provided, comprising computer instructions, which, when executed by a processor, perform any one of the above-mentioned risk analysis methods in a power grid operation process.

[0025] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium includes a stored program, wherein the program executes any one of the above-mentioned risk analysis methods during power grid operation.

[0026] According to another aspect of an embodiment of the present invention, a processor is provided, the processor being used to run a program, wherein when the program is run, the risk analysis method in the operation process of a power grid described in any one of the above is executed.

[0027] In an embodiment of the present invention, current grid state data of a target grid during operation is preprocessed to obtain preprocessed current grid state data; feature extraction is performed on the preprocessed current grid state data to obtain feature data of the target grid; the feature data is processed using a grid state data prediction model to obtain future grid state data of the target grid within a predetermined future period, wherein the grid state data prediction model is obtained by machine learning training using multiple sets of training data, each of the multiple sets of training data includes: sample feature data and sample future grid state data corresponding to the sample feature data; predetermined indicators of the target grid are determined using the future grid state data, wherein the predetermined indicators include at least: an expected load loss indicator, a grid state indicator, and an over-limit indicator, wherein the expected load loss indicator indicates the proportion of the load to be removed before and after a fault occurs in the target grid to the total load, and the over-limit indicator indicates the operating state of the target grid at a predetermined time; a risk index of the target grid is determined according to the indicator value corresponding to the predetermined indicator; and the risk level of the target grid is determined according to the risk index. Through the technical solution provided by the present invention, it is possible to predict the grid state data of the grid in the future based on the characteristic data extracted from the current grid state data, and then obtain the risk index of the grid based on the predicted grid state data analysis, and then determine its risk level according to the risk index, thereby achieving the technical effect of timely responding to changes in the grid operation state and dynamic changes in risks, improving the stability of the grid, and thus solving the technical problem that the method used for risk analysis of the grid operation process in the related technology can usually only analyze historical data, cannot respond to changes in the grid operation state and dynamic changes in risks in a timely manner, and has low reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0029] Figure 1 It is a hardware structure block diagram of a mobile terminal of a risk analysis method in a power grid operation process according to an embodiment of the present invention;

[0030] Figure 2 is a flow chart of a risk analysis method during power grid operation according to an embodiment of the present invention;

[0031] Figure 3 is a schematic diagram of a risk analysis device during power grid operation according to an embodiment of the present invention. DETAILED DESCRIPTION

[0032] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0033] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0034] As introduced in the background technology, the methods used in the related art for risk analysis of the power grid operation process can usually only analyze historical data, and cannot respond to changes in the power grid operation status and dynamic changes in risks in a timely manner, and have low reliability. In view of the above defects, in an embodiment of the present invention, a method and device for risk analysis in the power grid operation process, a computer-readable storage medium, a processor, and a computer program product are provided.

[0035] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0036] The method embodiments provided in the embodiments of the present invention can be executed in a mobile terminal, a computer terminal or a similar computing device. Taking running on a mobile terminal as an example, Figure 1 FIG. 1 is a hardware structure block diagram of a mobile terminal for a risk analysis method during power grid operation according to an embodiment of the present invention. Figure 1 As shown, the mobile terminal may include one or more ( Figure 1 Only one is shown in the figure) a processor 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA) and a memory 104 for storing data, wherein the mobile terminal may also include a transmission device 106 and an input / output device 108 for communication functions. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the mobile terminal. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations shown.

[0037] The memory 104 can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the risk analysis method during the operation of the power grid in the embodiment of the present invention. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, that is, the above method is implemented. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the mobile terminal via a network. Examples of the above-mentioned network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof. The transmission device 106 is used to receive or send data via a network. The above-mentioned specific examples of the network may include a wireless network provided by a communication provider of the mobile terminal. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, referred to as NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 may be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0038] According to an embodiment of the present invention, a method embodiment of a risk analysis method during power grid operation is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0039] Figure 2 is a flow chart of a risk analysis method during power grid operation according to an embodiment of the present invention. Figure 2 As shown, the method comprises the following steps:

[0040] Step S202 , preprocessing the current grid state data of the target grid during operation to obtain the preprocessed current grid state data.

[0041] Optionally, the current grid status data may include: real-time data such as grid load, voltage, frequency, and static data such as grid topology, equipment status, etc. In addition, it may also include: renewable energy power generation data, such as wind power generation and photovoltaic power generation. In addition, it may also include historical fault data, such as historical fault records and repair status of grid equipment.

[0042] It should be noted that the above current power grid status data may include but is not limited to the above contents, which will not be repeated here.

[0043] Optionally, the above preprocessing may include but is not limited to: missing value filling, noise filtering, dimensionality reduction, etc. The specific preprocessing method will be described below in conjunction with the specific implementation method. No further details will be given here.

[0044] The preprocessing method in this embodiment can facilitate the operation of the current power grid state data in subsequent calculation processes. For example, invalid data in the current power grid state data can be removed through noise filtering, thereby improving the data processing speed.

[0045] Step S204, extracting features from the preprocessed current power grid state data to obtain feature data of the target power grid.

[0046] In this embodiment, feature extraction may be performed on the preprocessed current power grid state data to obtain feature data of the target power grid, which facilitates subsequent prediction of power grid state data in future time periods.

[0047] Step S206, using the power grid state data prediction model to process the characteristic data to obtain the future power grid state data of the target power grid within a predetermined time period in the future, wherein the power grid state data prediction model is obtained through machine learning training using multiple sets of training data, and each of the multiple sets of training data includes: sample characteristic data and sample future power grid state data corresponding to the sample characteristic data.

[0048] In this embodiment, the feature data can be processed using a pre-trained grid state data prediction model to obtain the grid state data of the target grid in a period of time in the future, that is, the future grid state data. In this way, the future grid state data can be quickly predicted.

[0049] Step S208, using the future grid status data to determine the predetermined indicators of the target grid, wherein the predetermined indicators include at least: an expected load loss indicator, a grid status indicator, and an over-limit indicator. The expected load loss indicator indicates the proportion of the load that needs to be removed before and after a fault occurs in the target grid to the total load, and the over-limit indicator indicates the operating status of the target grid at a predetermined time.

[0050] Step S210: determining a risk index of a target power grid according to an indicator value corresponding to a predetermined indicator.

[0051] Step S212, determining the risk level of the target power grid according to the risk index.

[0052] In this embodiment, the risk index can be matched with the risk index / risk level mapping relationship to obtain the risk level. In addition, a risk level determination model can also be obtained through pre-training to use the risk level determination model to determine the risk level based on the risk index.

[0053] As can be seen from the above, in an embodiment of the present invention, the current grid state data of the target grid during operation is preprocessed to obtain the preprocessed current grid state data; the preprocessed current grid state data is feature extracted to obtain the feature data of the target grid; the feature data is processed using a grid state data prediction model to obtain the future grid state data of the target grid in a predetermined future period; the predetermined indicators of the target grid are determined using the future grid state data, wherein the predetermined indicators at least include: an expected load loss indicator, a grid state indicator, and an over-limit indicator, the expected load loss indicator indicates the proportion of the load required to be removed before and after the target grid fails to the total load, and the over-limit indicator indicates the operating state of the target grid at a predetermined time; the risk index of the target grid is determined according to the indicator value corresponding to the predetermined indicator; the risk level of the target grid is determined according to the risk index, so that the grid state data of the grid in the future period of time is predicted according to the feature data extracted from the current grid state data, and then the risk index of the grid is obtained based on the predicted grid state data. The purpose of determining its risk level according to the risk index achieves the technical effect of timely responding to changes in the grid operation state and dynamic changes in risks, and improves the stability of the grid.

[0054] The above-mentioned technical solution provided by the embodiment of the present invention solves the technical problem that the method used in the related technology for risk analysis of the power grid operation process can usually only analyze historical data, cannot respond to changes in the power grid operation status and dynamic changes in risks in a timely manner, and has low reliability.

[0055] According to the above embodiment of the present invention, before preprocessing the current grid status data of the target grid during operation to obtain the preprocessed current grid status data, the risk analysis method during the grid operation process also includes: when it is determined that the target grid enters the operating state, triggering the first sensor on each grid circuit in the target grid and the second sensor set at a predetermined position of each electrical equipment in the target grid to start, so as to collect the current grid status data of the target grid; using the Internet of Things device to obtain the current grid status data from the first sensor and the second sensor, wherein the current grid status data includes: operating parameters of the target grid, grid topology and equipment status, new energy power generation data in the target grid and historical fault data of the target grid.

[0056] Optionally, the first sensor and the second sensor may include but are not limited to: a current sensor, a voltage sensor, a mutual inductor, etc.

[0057] In this embodiment, before preprocessing the current grid state data, the current grid state data may be acquired first. For example, the state data on the grid circuit may be collected using an inductive component (i.e., a first sensor) disposed on the grid (i.e., the target grid) circuit, and the state data at each electrical device may be collected using an inductive component (i.e., a second sensor) disposed on each electrical device in the grid.

[0058] Next, various sensor data in the power grid can be collected through IoT devices to obtain current power grid status data.

[0059] According to the above-mentioned embodiment of the present invention, the current grid state data of the target grid during operation is preprocessed to obtain the preprocessed current grid state data, including at least one of the following: interpolating the missing values ​​in the current grid state data to obtain the current grid state data after interpolation; filtering the noise data in the current grid state data to obtain the filtered current grid state data; performing dimensionality reduction processing on the current grid state data to obtain the current grid state data after dimensionality reduction processing.

[0060] In this embodiment, after the power grid status data is automatically acquired through the power grid data acquisition technology, cubic spline interpolation is used to process missing values, VMD is used to perform data denoising, and then Autoencoder is used to perform data dimensionality reduction to obtain preprocessed power grid status data.

[0061] Among them, the above VMD (Variational Mode Decomposition) is a signal processing method used to remove noise from signals. Its principle is based on adaptive decomposition and modal filtering of signals. The above Autoencoder is an unsupervised learning neural network model, which can be used for feature extraction and dimensionality reduction of data. It consists of an encoder and a decoder. The encoder compresses the input data into a potential representation, and the decoder restores the potential representation to the original data. By training the autoencoder, an effective representation of the data can be learned, which can be used for tasks such as data compression, noise reduction, and feature extraction.

[0062] In the above-mentioned embodiment of the present invention, interpolation processing is performed on the missing values ​​in the current power grid state data to obtain the current power grid state data after interpolation processing, including: determining the interpolation step size of the interpolation processing according to the total amount of data of the current power grid state data; determining the curve coefficient of the spline curve used for the interpolation processing; determining the interpolation data of the missing values ​​according to the interpolation step size, the curve coefficient and the current power grid state data; and inserting the interpolation data into the current power grid state data according to the interpolation step size to obtain the current power grid state data after interpolation processing.

[0063] In this embodiment, the missing data in the various data collected by the power grid data collection technology can be interpolated using the cubic spline interpolation method to avoid information loss; the interpolation principle is: there are m data nodes (x1, y1)(x2, y2)…(x n ,y n ), where x represents the time of data collection, y represents the collected power grid status data value, and the calculation step length H i =x i+1 -x i (i=1,2,...n), calculate the coefficient A of the spline curve i ,B i ,C i ,D i , Finally, in each subinterval x i ≤x≤x i+1 , create the equation: G i (x) = A i +B i (xx i )+C i (xx i ) 2 +D i (xx i ) 3 , where: G i (x) is the interpolated data of the missing value at the i-th position.

[0064] The above interpolation processing can fill in the missing values ​​in the current power grid status data to ensure data integrity, thereby making the risk analysis results during the power grid operation more accurate.

[0065] In the above embodiment of the present invention, filtering the noise data in the current power grid state data to obtain the filtered current power grid state data may include: decomposing the state signal corresponding to the current power grid state data by using a variational mode decomposition method to obtain the modal function corresponding to the current power grid state data; denoising the modal function by using a modal filter to obtain the denoised state signal; and obtaining the filtered current power grid state data according to the denoised state signal.

[0066] In this embodiment, the VMD method can be used to remove noise in the interpolated data through a combination of adaptive decomposition and modal filtering, thereby improving the quality of the data while retaining the essential characteristics of the data; the VMD noise reduction steps here are: first, VMD decomposes the input signal into a series of inherent modal functions, each modal function represents a specific vibration mode; the decomposition process uses the variational principle to optimize the separation of each modal function and make it contain the local frequency information in the signal as much as possible; then VMD applies a modal filter to denoise each modal function, and the modal filter can selectively enhance or suppress specific frequency components in the modal function, thereby improving the signal quality while retaining the essential characteristics of the signal.

[0067] The denoising method here can remove invalid data in the power grid status data, reduce the data processing volume, and increase the data processing speed.

[0068] In the above embodiment of the present invention, the current power grid state data is subjected to dimensionality reduction processing to obtain the current power grid state data after dimensionality reduction processing, including: inputting the current power grid state data into a target encoder, so as to use the target encoder to perform dimensionality reduction processing on the current power grid state data, so as to obtain the current power grid state data after dimensionality reduction processing, wherein the target encoder is an encoder obtained by the following training of the original encoder: inputting sample power grid state data into the original encoder, so as to use the original encoder to process the sample power grid state data, so as to obtain processed sample coded data; inputting the processed sample coded data into a decoder, so as to use the decoder to process the processed sample coded data, so as to obtain decoded sample coded data; using a loss function to calculate the data difference between the processed sample coded data and the decoded sample coded data; when it is determined based on the data difference that the training convergence condition is met, the target encoder is obtained.

[0069] In this embodiment, the denoised data is compressed and reconstructed through Autoencoder to reduce the data dimension, extract important information from the data, and improve the model prediction speed. The dimensionality reduction process using Autoencoder is as follows: define a neural network model as an encoder, and the input layer of the encoder receives the high-dimensional representation of the data processed in step 1.2; define another neural network model as a decoder, and the input layer of the decoder receives the output of the encoder. Use the reconstruction error as the loss function to measure the difference between the output of the decoder and the original data. The encoder and decoder are jointly trained by the back propagation algorithm and the gradient descent algorithm to minimize the loss function. After the training is completed, the model of the encoder part can be used to map the original data to a low-dimensional space to achieve data compression and reconstruction.

[0070] The above-mentioned dimensionality reduction method can reduce the dimension of the power grid status data, thereby improving the data processing speed, and also improving the speed of obtaining the power grid risk analysis results.

[0071] According to the above-mentioned embodiment of the present invention, the characteristic data is processed by using a power grid state data prediction model to obtain the future power grid state data of the target power grid within a predetermined time period in the future, including: using a convolutional neural network to extract features of the current power grid state data to convert the time series data corresponding to the current power grid state data into a current power grid state data sequence with spatial characteristics; obtaining time dependency data in the current power grid state data based on the current power grid state data sequence; formatting the time dependency data to obtain time dependency data identified by an integrated learning algorithm; and using an integrated learning algorithm to predict the future power grid state data based on the time dependency data and the current power grid state data sequence with spatial characteristics.

[0072] In this embodiment, a convolutional neural network (CNN) can be used to extract features from the power grid state data, that is, CNN is mainly used for feature extraction in sequence prediction, and CNN extracts spatial features of data through convolutional layers and pooling layers. BiLSTM is used for modeling power grid sequence data. The BiLSTM structure contains forward and backward LSTM units, which can better capture the context information of power grid data. XGBoost is a gradient boosting algorithm that improves prediction performance by integrating multiple decision trees, and performs well in predicting large data sets such as power grid operation status.

[0073] That is, in this embodiment, the grid state data is predicted by the CNN-BiLSTM-XGB algorithm. The CNN-BiLSTM-XGB algorithm is a machine learning model combination that combines a convolutional neural network (CNN), a bidirectional long short-term memory network (BiLSTM), and XGBoost (XGB). This combination model is usually used to process sequence data and perform prediction or classification tasks based on this.

[0074] Specifically, the specific process of using CNN-BiLSTM-XGB for prediction is as follows: 1) Use CNN to extract features of power grid data and convert time series data into a representation with spatial features; 2) Input the feature sequence extracted by CNN into BiLSTM for sequence modeling to capture the temporal dependency in the data; 3) Convert the feature sequence output by BiLSTM into a format acceptable to XGBoost, and then use XGBoost to predict the operation data of the power grid in the future.

[0075] According to the above embodiment of the present invention, determining the risk index of the target power grid according to the index value corresponding to the predetermined index includes: determining the index value one corresponding to the over-limit index by a first formula, wherein the first formula is: M0 represents the index value 1, n represents the number of buses in the target power grid, k represents the kth bus in the target power grid, Indicates the voltage over-limit situation of the kth busbar; the second indicator value corresponding to the expected load loss indicator is determined by the second formula, where the second formula is: M Ll =∑ c∈M p c *L c , M Ll Indicates the index value 2, L c Indicates the load lost by the target power grid during operation, p c represents the risk operation probability of the target power grid in the cth state; the indicator value three corresponding to the power grid state indicator is determined by the third formula, where the third formula is: M Gs =∑ c∈Gsi p c , M Gs represents the index value three, Gsi represents the set of target power grids in the predetermined operating state i; the risk index is determined according to the index value one, the index value two and the index value three.

[0076] As can be seen from the above, in the embodiment of the present invention, the power grid risk assessment index can be divided into a voltage over-limit index, an expected load loss index and a power grid state index. Among them, the voltage over-limit index represents the operating state of the power grid at a certain moment. By dividing these three indicators, the impact of the accident on the power grid can be quantified.

[0077] The voltage over-limit index indicates the degree to which the voltage of each bus in the power grid exceeds the rated voltage, and its calculation formula is the first formula above. It can be determined by the following formula: Where: U k is the bus voltage, U up is the upper limit of bus voltage, U low The lower limit of bus voltage.

[0078] The above-mentioned expected load loss index indicates the degree to which the load that needs to be removed by linear optimal load calculation before and after the fault accounts for the total load. The calculation formula is the same as the second formula above.

[0079] In the embodiment of the present invention, the power grid status index can be divided into a normal state, an over-limit state and an accident state. The power grid status index calculation formula is the third formula above.

[0080] According to the above embodiment of the present invention, determining the risk index according to indicator value one, indicator value two and indicator value three includes: obtaining the weight values ​​of indicator value one, indicator value two and indicator value three respectively; determining the risk index according to indicator value one, indicator value two and indicator value three and their respective corresponding weight values.

[0081] In this embodiment, the entropy method can be used to calculate the expected load loss index, grid status index, and over-limit index weights. The entropy method weight calculation formula is as follows: (When f ij =0, f ij lnf ij =0), where m is the evaluation year, n is the evaluation index, and w j is the weight value of the jth indicator, H j is the entropy of the jth index, f ij Represents the original value of the evaluation matrix, A ij is the standardized evaluation data matrix.

[0082] Next, the grid operation risk index is calculated based on the expected load loss index, grid state index, and over-limit index of the grid. The calculation formula of the operation risk index λ is: λ=M Ll *W Ll +M Gs *W Gs +M O *W O , where: λ is the power grid operation risk index, W Ll is the expected load loss index weight, W Gs is the power grid status indicator weight, W O is the weight of the over-limit indicator.

[0083] The above technical solution provided in the embodiment of the present invention is described below in conjunction with Table 1, wherein Table 1 shows the power grid operation risk index and risk level:

[0084] Table 1

[0085]

[0086]

[0087] Table 1 above records the corresponding relationship between each risk index and risk level, as well as the risk assessment and possible consequences corresponding to different risk levels. Through this table, the risk level corresponding to different risk indicators and the risk assessment results can be intuitively determined. At the same time, the possible consequences corresponding to different risk assessment results can also be obtained, which is convenient for subsequent maintenance personnel to conduct timely maintenance on the power grid and determine the safety of the power grid.

[0088] Through the technical solution provided by the above embodiment of the present invention, the grid state data is first automatically acquired through the grid data acquisition technology, the missing values ​​are processed by cubic spline interpolation, the data is denoised by VMD, and then the data dimension is reduced by Autoencoder; the CNN-BiLSTM-XGB model is used to extract data features to predict the operation state of the grid in the future; the grid operation state data obtained by the model prediction is used to divide the grid risk assessment index into expected load loss index, grid state index, and over-limit index and calculate; the grid operation risk is graded according to the results of the risk assessment index. Comprehensive consideration of multiple factors such as grid state and new energy forecast data can more comprehensively assess the grid operation risk. The use of data-driven methods for analysis can more accurately assess the grid operation risk, avoid the shortcomings of relying on empirical rules in traditional methods, and use real-time data for predictive analysis, so that potential risks in grid operation can be discovered in a timely manner and measures can be taken in a timely manner.

[0089] That is, the technical solution provided by the embodiment of the present invention comprehensively considers multiple factors, adopts appropriate data processing methods, and advanced prediction models, and can more comprehensively evaluate the risk of power grid operation. At the same time, the use of data-driven methods for analysis can more accurately evaluate the risk of power grid operation, avoiding the shortcomings of traditional methods that rely on empirical rules, and can help managers to promptly discover the risk of power grid operation, which helps power grid operation managers to take timely measures to ensure the safety and stability of power grid operation.

[0090] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0091] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0092] According to an embodiment of the present invention, there is also provided a risk analysis device in a power grid operation process for implementing the risk analysis method in a power grid operation process. Figure 3 is a schematic diagram of a risk analysis device during power grid operation according to an embodiment of the present invention. Figure 3 As shown, the device includes: a preprocessing module 301, a feature extraction module 303, a processing module 305, a first determination module 307, a second determination module 309 and a third determination module 311. The risk analysis device in the operation process of the power grid is described below.

[0093] The preprocessing module 301 is used to preprocess the current grid state data of the target grid during operation to obtain the preprocessed current grid state data.

[0094] The feature extraction module 303 is used to extract features from the preprocessed current power grid state data to obtain feature data of the target power grid.

[0095] The processing module 305 is used to process the characteristic data using the power grid state data prediction model to obtain the future power grid state data of the target power grid within a predetermined time period in the future, wherein the power grid state data prediction model is obtained through machine learning training using multiple sets of training data, and each of the multiple sets of training data includes: sample characteristic data and sample future power grid state data corresponding to the sample characteristic data.

[0096] The first determination module 307 is used to determine the predetermined indicators of the target power grid using the future power grid status data, wherein the predetermined indicators include at least: an expected load loss indicator, a power grid status indicator and an over-limit indicator. The expected load loss indicator indicates the proportion of the load that needs to be removed before and after the target power grid fails to the total load, and the over-limit indicator indicates the operating status of the target power grid at a predetermined time.

[0097] The second determination module 309 is used to determine the risk index of the target power grid according to the indicator value corresponding to the predetermined indicator.

[0098] The third determination module 311 is used to determine the risk level of the target power grid according to the risk index.

[0099] It should be noted here that the above-mentioned preprocessing module 301, feature extraction module 303, processing module 305, first determination module 307, second determination module 309 and third determination module 311 correspond to steps S202 to S212 in the above-mentioned embodiments, and the six modules and the corresponding steps have the same instances and application scenarios implemented, but are not limited to the contents disclosed in the above-mentioned embodiments.

[0100] As can be seen from the above, in the scheme recorded in the above-mentioned embodiment of the present invention, first, the preprocessing module can be used to preprocess the current grid state data of the target grid during operation to obtain the preprocessed current grid state data; then, the feature extraction module can be used to extract features from the preprocessed current grid state data to obtain the feature data of the target grid; then, the processing module can be used to process the feature data using the grid state data prediction model to obtain the future grid state data of the target grid within a predetermined time period in the future, wherein the grid state data prediction model is obtained by machine learning training using multiple sets of training data, and each set of the multiple sets of training data includes: sample feature data and sample future grid state data corresponding to the sample feature data; secondly, the first determination module can be used to determine the predetermined time period of the target grid using the future grid state data. Indicators, wherein the predetermined indicators at least include: an expected load loss indicator, a power grid state indicator and an over-limit indicator, the expected load loss indicator indicates the proportion of the load that needs to be removed before and after a fault occurs in the target power grid to the total load, and the over-limit indicator indicates the operating state of the target power grid at a predetermined time; a second determination module can also be used to determine the risk index of the target power grid according to the indicator value corresponding to the predetermined indicator; finally, a third determination module can be used to determine the risk level of the target power grid according to the risk index, thereby realizing the prediction of the power grid state data of the power grid in a future period of time based on the characteristic data extracted from the current power grid state data, and then obtaining the risk index of the power grid based on the predicted power grid state data analysis, and then determining its risk level according to the risk index, thereby achieving the technical effect of timely responding to changes in the power grid operating state and dynamic changes in risks, and improving the stability of the power grid.

[0101] The above-mentioned technical solution provided by the embodiment of the present invention solves the technical problem that the method used in the related technology for risk analysis of the power grid operation process can usually only analyze historical data, cannot respond to changes in the power grid operation status and dynamic changes in risks in a timely manner, and has low reliability.

[0102] In an optional embodiment, the risk analysis device during the operation of the power grid also includes: a collection module, which is used to preprocess the current power grid status data of the target power grid during the operation process, and before obtaining the preprocessed current power grid status data, when it is determined that the target power grid enters the operating state, trigger the first sensor on each power grid circuit in the target power grid and the second sensor set at a predetermined position of each electrical equipment in the target power grid to start, so as to collect the current power grid status data of the target power grid; an acquisition module, which is used to use the Internet of Things device to obtain the current power grid status data from the first sensor and the second sensor, wherein the current power grid status data includes: the operating parameters of the target power grid, the power grid topology and the equipment status, the new energy power generation data in the target power grid and the historical fault data of the target power grid.

[0103] In an optional embodiment, the preprocessing module includes at least one of the following: an interpolation unit, used to interpolate missing values ​​in the current power grid status data to obtain the current power grid status data after interpolation; a filtering unit, used to filter noise data in the current power grid status data to obtain the current power grid status data after filtering; a dimensionality reduction unit, used to perform dimensionality reduction on the current power grid status data to obtain the current power grid status data after dimensionality reduction.

[0104] In an optional embodiment, the interpolation unit includes: a first determination subunit, used to determine the interpolation step size of the interpolation processing according to the total amount of data of the current power grid status data; a second determination subunit, used to determine the curve coefficient of the spline curve used for the interpolation processing; a third determination subunit, used to determine the interpolation data of the missing value according to the interpolation step size, the curve coefficient and the current power grid status data; an insertion subunit, used to insert the interpolation data into the current power grid status data according to the interpolation step size to obtain the current power grid status data after interpolation processing.

[0105] In an optional embodiment, the filtering unit includes: a decomposition subunit, which is used to decompose the state signal corresponding to the current power grid state data using a variational mode decomposition method to obtain the modal function corresponding to the current power grid state data; a denoising subunit, which is used to denoise the modal function using a modal filter to obtain the denoised state signal; and a first acquisition subunit, which is used to obtain the filtered current power grid state data according to the denoised state signal.

[0106] In an optional embodiment, the dimension reduction unit includes: a dimension reduction sub-unit, used to input the current power grid state data into a target encoder, so as to use the target encoder to perform dimension reduction processing on the current power grid state data, so as to obtain the current power grid state data after dimension reduction processing, wherein the target encoder is an encoder obtained by the following training of the original encoder: inputting the sample power grid state data into the original encoder, so as to use the original encoder to process the sample power grid state data, so as to obtain processed sample coded data; inputting the processed sample coded data into a decoder, so as to use the decoder to process the processed sample coded data, so as to obtain decoded sample coded data; using the loss function to calculate the data difference between the processed sample coded data and the decoded sample coded data; when it is determined based on the data difference that the training convergence condition is met, the target encoder is obtained.

[0107] In an optional embodiment, the processing module includes: a feature extraction unit, which is used to extract features of current power grid status data using a convolutional neural network to convert time series data corresponding to the current power grid status data into a current power grid status data sequence with spatial features; an acquisition unit, which is used to acquire time dependency data in the current power grid status data based on the current power grid status data sequence; a format conversion unit, which is used to perform format conversion on the time dependency data to obtain time dependency data recognized by an integrated learning algorithm; and a prediction unit, which is used to predict future power grid status data based on the time dependency data and the current power grid status data sequence with spatial features using an integrated learning algorithm.

[0108] In an optional embodiment, the second determination module includes: a first determination unit, configured to determine an index value 1 corresponding to the over-limit index by using a first formula, wherein the first formula is: , M0 represents the index value 1, n represents the number of buses in the target power grid, k represents the kth bus in the target power grid, Indicates the voltage over-limit situation of the kth bus; a second determination unit, used to determine the index value 2 corresponding to the expected load loss index through a second formula, wherein the second formula is: M Ll =∑ c∈M p c *L c , M Ll Indicates the index value 2, L c Indicates the load lost by the target power grid during operation, p c represents the risk operation probability of the target power grid in the cth state; a third determination unit is used to determine the indicator value three corresponding to the power grid state indicator through a third formula, wherein the third formula is: M Gs =∑ c∈Gsi p c , M Gs represents index value three, Gsi represents the set of target power grids in a predetermined operating state i; a fourth determination unit is used to determine the risk index according to index value one, index value two and index value three.

[0109] In an optional embodiment, the fourth determination unit includes: a second acquisition subunit, used to obtain the weight values ​​of indicator value one, indicator value two and indicator value three respectively; a fourth determination subunit, used to determine the risk index based on indicator value one, indicator value two and indicator value three and their respective corresponding weight values.

[0110] According to another aspect of an embodiment of the present invention, a processor is further provided, and the processor is used to run a program, wherein when the program is run, any one of the above-mentioned risk analysis methods in the operation process of a power grid is executed.

[0111] According to another aspect of an embodiment of the present invention, a computer program product is provided, including computer instructions, which, when executed by a processor, perform any one of the above-mentioned risk analysis methods in a power grid operation process.

[0112] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium comprising a stored program, wherein the program executes any one of the above-mentioned risk analysis methods during power grid operation.

[0113] Optionally, in this embodiment, the computer-readable storage medium may be located in any one of the computer terminals in a computer terminal group in a computer network, or in any one of the communication devices in a communication device group.

[0114] Optionally, in this embodiment, the computer-readable storage medium is configured to store program codes for executing the following steps: preprocessing current grid state data of the target grid during operation to obtain preprocessed current grid state data; extracting features from the preprocessed current grid state data to obtain feature data of the target grid; processing the feature data using a grid state data prediction model to obtain future grid state data of the target grid within a predetermined future period of time, wherein the grid state data prediction model is obtained through machine learning training using multiple sets of training data, each of the multiple sets of training data including: sample feature data and sample future grid state data corresponding to the sample feature data; determining predetermined indicators of the target grid using the future grid state data, wherein the predetermined indicators include at least: an expected load loss indicator, a grid state indicator, and an over-limit indicator, wherein the expected load loss indicator indicates the proportion of the load to be removed before and after a fault occurs in the target grid to the total load, and the over-limit indicator indicates the operating status of the target grid at a predetermined time; determining a risk index of the target grid according to the indicator value corresponding to the predetermined indicator; and determining the risk level of the target grid according to the risk index.

[0115] Optionally, in this embodiment, the computer-readable storage medium is configured to store program code for executing the following steps: before preprocessing the current grid status data of the target power grid during operation to obtain the preprocessed current grid status data, when it is determined that the target power grid enters an operating state, triggering a first sensor on each grid circuit in the target power grid and a second sensor arranged at a predetermined position of each electrical equipment in the target power grid to start, so as to collect the current grid status data of the target power grid; using an Internet of Things device to obtain the current grid status data from the first sensor and the second sensor, wherein the current grid status data includes: operating parameters, grid topology and equipment status of the target power grid, new energy power generation data in the target power grid and historical fault data of the target power grid.

[0116] Optionally, in this embodiment, the computer-readable storage medium is configured to store program codes for executing the following steps: interpolating missing values ​​in the current power grid status data to obtain the current power grid status data after interpolation; filtering noise data in the current power grid status data to obtain the current power grid status data after filtering; performing dimensionality reduction processing on the current power grid status data to obtain the current power grid status data after dimensionality reduction processing.

[0117] Optionally, in this embodiment, the computer-readable storage medium is configured to store program codes for executing the following steps: determining an interpolation step size for interpolation processing based on the total amount of data of the current power grid status data; determining a curve coefficient of a spline curve for interpolation processing; determining interpolation data for missing values ​​based on the interpolation step size, the curve coefficient and the current power grid status data; and inserting the interpolation data into the current power grid status data according to the interpolation step size to obtain the current power grid status data after interpolation processing.

[0118] Optionally, in this embodiment, the computer-readable storage medium is configured to store program codes for executing the following steps: decomposing a state signal corresponding to the current power grid state data using a variational modal decomposition method to obtain a modal function corresponding to the current power grid state data; denoising the modal function using a modal filter to obtain a denoised state signal; and obtaining the filtered current power grid state data based on the denoised state signal.

[0119] Optionally, in this embodiment, the computer-readable storage medium is configured to store program codes for executing the following steps: inputting current power grid state data into a target encoder, so as to use the target encoder to perform dimensionality reduction processing on the current power grid state data, so as to obtain the current power grid state data after dimensionality reduction processing, wherein the target encoder is an encoder obtained by the following training of the original encoder: inputting sample power grid state data into the original encoder, so as to use the original encoder to process the sample power grid state data, so as to obtain processed sample coded data; inputting the processed sample coded data into a decoder, so as to use the decoder to process the processed sample coded data, so as to obtain decoded sample coded data; using a loss function to calculate the data difference between the processed sample coded data and the decoded sample coded data; when it is determined that the training convergence condition is met based on the data difference, the target encoder is obtained.

[0120] Optionally, in this embodiment, the computer-readable storage medium is configured to store program code for executing the following steps: using a convolutional neural network to extract features of the current power grid state data to convert the time series data corresponding to the current power grid state data into a current power grid state data sequence with spatial characteristics; obtaining time dependency data in the current power grid state data based on the current power grid state data sequence; formatting the time dependency data to obtain time dependency data identified by an integrated learning algorithm; and using an integrated learning algorithm to predict future power grid state data based on the time dependency data and the current power grid state data sequence with spatial characteristics.

[0121] Optionally, in this embodiment, the computer-readable storage medium is configured to store a program code for executing the following steps: determining an index value 1 corresponding to the over-limit index by a first formula, wherein the first formula is: M0 represents the index value 1, n represents the number of buses in the target power grid, k represents the kth bus in the target power grid, Indicates the voltage over-limit situation of the kth busbar; the second indicator value corresponding to the expected load loss indicator is determined by the second formula, where the second formula is: M Ll =∑ c∈M p c *L c , M Ll Indicates the index value 2, L c Indicates the load lost by the target power grid during operation, p c represents the risk operation probability of the target power grid in the cth state; the indicator value three corresponding to the power grid state indicator is determined by the third formula, where the third formula is: M Gs =∑ c∈Gsi p c , M Gs represents the index value three, Gsi represents the set of target power grids in the predetermined operating state i; the risk index is determined according to the index value one, the index value two and the index value three.

[0122] Optionally, in this embodiment, the computer-readable storage medium is configured to store program code for executing the following steps: respectively obtaining weight values ​​of indicator value one, indicator value two and indicator value three; determining a risk index based on indicator value one, indicator value two and indicator value three and their respective corresponding weight values.

[0123] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0124] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0125] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0126] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A risk analysis method in the operation process of a power grid, characterized in that: include: Preprocessing the current grid state data of the target grid during operation to obtain the preprocessed current grid state data; Performing feature extraction on the preprocessed current power grid state data to obtain feature data of the target power grid; The characteristic data is processed using a power grid state data prediction model to obtain future power grid state data of the target power grid within a predetermined future period, wherein the power grid state data prediction model is obtained by machine learning training using multiple sets of training data, and each of the multiple sets of training data includes: sample characteristic data and sample future power grid state data corresponding to the sample characteristic data; Determine a predetermined index of the target power grid by using the future power grid state data, wherein the predetermined index at least includes: an expected load loss index, a power grid state index, and an over-limit index, wherein the expected load loss index indicates the proportion of the load required to be removed before and after the target power grid fails to the total load, and the over-limit index indicates the operating state of the target power grid at a predetermined time; Determining a risk index of the target power grid according to an indicator value corresponding to the predetermined indicator; The risk level of the target power grid is determined according to the risk index.

2. The risk analysis method in the operation process of the power grid according to claim 1 is characterized in that: Before preprocessing the current grid state data of the target grid during operation to obtain the preprocessed current grid state data, the method further includes: When it is determined that the target power grid enters the operating state, triggering a first sensor on each power grid circuit in the target power grid and a second sensor arranged at a predetermined position of each electrical device in the target power grid to start, so as to collect the current power grid state data of the target power grid; The current grid status data is obtained from the first sensor and the second sensor using an Internet of Things device, wherein the current grid status data includes: operating parameters, grid topology and equipment status of the target grid, new energy power generation data in the target grid and historical fault data of the target grid.

3. The risk analysis method in the operation process of the power grid according to claim 1 is characterized in that: Preprocessing the current grid state data of the target grid during operation to obtain the preprocessed current grid state data includes at least one of the following: Performing interpolation processing on missing values ​​in the current power grid state data to obtain the current power grid state data after interpolation processing; Filtering the noise data in the current power grid state data to obtain the filtered current power grid state data; Performing dimension reduction processing on the current power grid state data to obtain the current power grid state data after dimension reduction processing.

4. The risk analysis method in the operation process of the power grid according to claim 3 is characterized in that: Interpolating missing values ​​in the current power grid state data to obtain the current power grid state data after interpolation, including: Determining the interpolation step size of the interpolation processing according to the total amount of data of the current power grid state data; Determining curve coefficients of a spline curve used for the interpolation process; Determine interpolation data of the missing value according to the interpolation step size, the curve coefficient and the current grid state data; The interpolation data is inserted into the current power grid state data according to the interpolation step length to obtain the current power grid state data after interpolation processing.

5. The risk analysis method in the operation process of the power grid according to claim 3 is characterized in that: Filtering the noise data in the current power grid state data to obtain the filtered current power grid state data includes: Decomposing the state signal corresponding to the current power grid state data by using a variational mode decomposition method to obtain a mode function corresponding to the current power grid state data; Performing denoising processing on the modal function using a modal filter to obtain the denoised state signal; The current power grid state data after filtering is obtained according to the state signal after denoising.

6. The risk analysis method during power grid operation according to claim 3 is characterized in that: Performing dimensionality reduction processing on the current power grid state data to obtain the current power grid state data after dimensionality reduction processing, including: The current power grid state data is input into a target encoder so as to use the target encoder to perform dimensionality reduction processing on the current power grid state data so as to obtain the current power grid state data after dimensionality reduction processing, wherein the target encoder is an encoder obtained by the following training of the original encoder: the sample power grid state data is input into the original encoder so as to use the original encoder to process the sample power grid state data so as to obtain processed sample coded data; the processed sample coded data is input into a decoder so as to use the decoder to process the processed sample coded data so as to obtain decoded sample coded data; the data difference between the processed sample coded data and the decoded sample coded data is calculated using a loss function; when it is determined based on the data difference that the training convergence condition is met, the target encoder is obtained.

7. The risk analysis method in the operation process of the power grid according to claim 1 is characterized in that: The characteristic data is processed using a power grid state data prediction model to obtain future power grid state data of the target power grid within a predetermined future period, including: Using a convolutional neural network to perform feature extraction on the current power grid state data, so as to convert the time series data corresponding to the current power grid state data into a current power grid state data sequence with spatial features; Acquire time dependency data in the current power grid state data based on the current power grid state data sequence; Performing format conversion on the time dependency data to obtain the time dependency data identified by the ensemble learning algorithm; The future power grid state data is obtained by predicting the time dependency data and the current power grid state data sequence with spatial characteristics using the integrated learning algorithm.

8. The risk analysis method in the operation process of the power grid according to claim 1 is characterized in that: Determining the risk index of the target power grid according to the indicator value corresponding to the predetermined indicator includes: The index value 1 corresponding to the over-limit index is determined by a first formula, wherein the first formula is: M0 represents the index value 1, n represents the number of buses in the target power grid, k represents the kth bus in the target power grid, Indicates the voltage exceeding the limit of the kth bus; The second index value corresponding to the expected load loss index is determined by a second formula, wherein the second formula is: M Ll Indicates the index value 2, L c represents the load lost by the target power grid during operation, p c represents the risk operation probability of the target power grid in the cth state; The indicator value three corresponding to the power grid status indicator is determined by a third formula, wherein the third formula is: M Gs represents the index value three, Gsi represents the set of the target power grid being in a predetermined operating state i; The risk index is determined according to the indicator value one, the indicator value two, and the indicator value three.

9. The risk analysis method during power grid operation according to claim 8, characterized in that: Determining the risk index according to the index value 1, the index value 2, and the index value 3 includes: Obtaining weight values ​​of the index value one, the index value two, and the index value three respectively; The risk index is determined according to the index value one, the index value two, and the index value three and the weight values ​​corresponding to each of them.

10. A risk analysis device during power grid operation, characterized in that: include: A preprocessing module, used to preprocess the current grid state data of the target grid during operation to obtain the preprocessed current grid state data; A feature extraction module, used for extracting features from the preprocessed current power grid state data to obtain feature data of the target power grid; A processing module, used for processing the characteristic data using a power grid state data prediction model to obtain future power grid state data of the target power grid within a predetermined future period, wherein the power grid state data prediction model is obtained by machine learning training using multiple sets of training data, and each of the multiple sets of training data includes: sample characteristic data and sample future power grid state data corresponding to the sample characteristic data; A first determination module is used to determine a predetermined index of the target power grid using the future power grid state data, wherein the predetermined index at least includes: an expected load loss index, a power grid state index, and an over-limit index, wherein the expected load loss index indicates the proportion of the load required to be removed before and after the target power grid fails to the total load, and the over-limit index indicates the operating state of the target power grid at a predetermined time; A second determination module, configured to determine a risk index of the target power grid according to an indicator value corresponding to the predetermined indicator; The third determination module is used to determine the risk level of the target power grid according to the risk index.

Images