Electric power dispatching risk early warning method based on fused electric power data

Abstract

An electric power dispatching risk early warning method based on fused electric power data comprises the steps that initial electric power data are acquired, and the initial electric power data at least comprise power grid operation data, user power consumption data, meteorological data and equipment state data; performing data fusion on the initial power data to construct a power data matrix; inputting the power data matrix into a risk assessment model to obtain a risk assessment result; wherein the risk assessment model is used for extracting risk features in the power data matrix from multiple dimensions and obtaining a risk assessment result based on the risk features, and the risk assessment model is also used for dynamically allocating weights of the risk features. Therefore, various influence factors can be comprehensively considered, and the accuracy of power dispatching risk management is improved.

Description

Technical Field

[0001] This application belongs to the fields of power system technology and computer technology, and in particular relates to a power dispatch risk early warning method based on fused power data. Background Technology

[0002] Power dispatch risk management refers to a technical solution that uses algorithms and technologies to monitor and provide early warnings of potential risks in the power system, and leverages relevant hardware and storage media to achieve intelligent management. Its core objective is to improve the efficiency and security of power dispatch.

[0003] However, in the existing technology, due to the influence of various factors on power dispatch, the existing power dispatch risk management strategies cannot fully consider these influencing factors, resulting in low accuracy of risk management. Summary of the Invention

[0004] The technical problem to be solved by this application is how to provide a power dispatch risk early warning method based on fused power data, which can comprehensively consider multiple influencing factors and improve the accuracy of power dispatch risk management.

[0005] To address the aforementioned problems, in a first aspect, this application provides a power dispatch risk early warning method based on fused power data. The method includes: acquiring initial power data, which at least includes grid operation data, user electricity consumption data, meteorological data, and equipment status data; fusing the initial power data to construct a power data matrix; and inputting the power data matrix into a risk assessment model to obtain a risk assessment result. The risk assessment model is used to extract risk features from the power data matrix from multiple dimensions and obtain the risk assessment result based on the risk features. The risk assessment model is also used to dynamically allocate weights to the risk features.

[0006] Optionally, the risk assessment model dynamically allocates the weights of the risk features through the following steps: standardizing the power data matrix according to a sliding time window; calculating the Mahalanobis distance between different risk features based on the standardized power data matrix; and dynamically allocating the weights of the risk features based on the Mahalanobis distance.

[0007] Optionally, the risk assessment model dynamically allocates the weights of the risk characteristics through the following steps: determining the core variables that affect the weight allocation; tracking the data fluctuations of the core variables in real time to obtain the change patterns; and setting the update frequency of the weight allocation based on the change patterns.

[0008] Optionally, the risk assessment model is constructed based on random forest and Bayesian network, wherein the random forest is used to process high-dimensional data in the power data matrix to identify the correlation between the nonlinear risk features, and the Bayesian network is used to update the risk assessment results in conjunction with the real-time power data matrix.

[0009] Optionally, the parameters of the risk assessment model can be optimized through a transfer learning mechanism.

[0010] Optionally, the data in the power data matrix includes the following dimensions: time dimension, spatial dimension, and feature channel.

[0011] Optionally, the spatial dimension of the data in the power data matrix is ​​generated based on the following steps: dividing the power grid coverage area into a dynamic spatial grid; mapping the initial power data to the dynamic spatial grid based on spatial association rules; and spatially fusing the initial power data.

[0012] Optionally, the time dimension of the data in the power data matrix is ​​generated based on the following steps: interpolating multiple initial power data according to a sliding window to generate multiple consecutive intermediate data; downsampling the high-frequency intermediate data, and / or upsampling the low-frequency intermediate data; and filling in the missing values ​​of the intermediate data.

[0013] In a second aspect, this application also provides a power dispatch risk early warning device based on fused power data, comprising: a data acquisition module for acquiring initial power data, wherein the initial power data includes at least grid operation data, user electricity consumption data, meteorological data, and equipment status data; a data fusion module for fusing the initial power data to construct a power data matrix; and an evaluation module for inputting the power data matrix into a risk evaluation model to obtain a risk evaluation result; wherein the risk evaluation model is used to extract risk features from the power data matrix from multiple dimensions and obtain a risk evaluation result based on the risk features, and the risk evaluation model is also used to dynamically allocate the weights of the risk features.

[0014] In a third aspect, this application also provides a computer device including a processor and a storage device, the storage device being adapted to store a plurality of program codes, the program codes being adapted to be loaded by the processor and executed to perform the method described in any of the preceding claims.

[0015] In a fourth aspect, this application also provides a storage medium storing a plurality of program codes adapted to be loaded and run by a processor to perform the methods described in any of the preceding claims.

[0016] Compared with the prior art, the technical solution of this application embodiment has the following beneficial effects:

[0017] The power dispatch risk early warning method based on fused power data in this application acquires multiple initial power data sets and performs fusion analysis on these data to conduct a more comprehensive risk analysis of the power data. Based on dynamic risk assessment, it provides risk early warning for power grid operation and adjusts the power grid's operating status. In this scheme, the weights of each risk feature in the model can be adjusted in real time based on changes in the power data matrix, making the risk assessment results more adaptable to the complex changes in power data and avoiding the influence of coupling relationships between related data on the accuracy of the risk assessment results. Therefore, it can comprehensively consider multiple influencing factors and improve the accuracy of power dispatch risk management. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating a power dispatch risk early warning method based on fused power data, according to an embodiment of this application.

[0019] Figure 2 This is a schematic diagram illustrating a process for dynamically allocating the weights of the risk features according to an embodiment of this application;

[0020] Figure 3 This is a schematic diagram illustrating another process for dynamically allocating the weights of the risk features according to an embodiment of this application;

[0021] Figure 4 This is a flowchart illustrating a method for aligning the time dimension of data in a power data matrix according to an embodiment of this application.

[0022] Figure 5 This is a schematic diagram of the structure of a power dispatch risk early warning device based on fused power data, according to an embodiment of this application. Detailed Implementation

[0023] As mentioned in the background section, existing AC motors generate a large amount of ineffective energy consumption during use, resulting in poor energy utilization efficiency.

[0024] To address the aforementioned issues, this invention proposes a power dispatch risk early warning method based on fused power data, which can be executed by the control terminal of an AC motor.

[0025] To further illustrate the technical means and effects adopted by this application to achieve the intended application objectives, the following detailed description, in conjunction with the accompanying drawings and preferred embodiments, explains the specific implementation methods and effects of the power dispatch risk early warning method based on fused power data proposed in this application.

[0026] According to the first aspect of this application, please refer to Figure 1 , Figure 1This is a flowchart illustrating a power dispatch risk early warning method based on fused power data. The method includes the following steps S101, S102, and S103, which are explained in detail below.

[0027] Step S101: Obtain initial power data, which includes at least grid operation data, user electricity consumption data, meteorological data, and equipment status data.

[0028] Among them, the power grid operation data refers to data from the power grid side, such as one or more of the following: voltage, current, frequency, substation load, etc. Power grid operation data can come from the SCADA monitoring system.

[0029] User electricity consumption data refers to data from the power grid's electricity consumption side, such as historical records of residential electricity consumption and current load levels. User electricity consumption data can be obtained from electricity consumption data collection terminals.

[0030] Meteorological data refers to weather conditions near the power grid, used to reflect the grid's operating environment. This includes data such as temperature, wind speed, lightning information, and the impact of sunlight intensity and wind speed on photovoltaic and wind power generation. Meteorological data can be accessed via an interface from the National Meteorological Center.

[0031] Equipment status data refers to data related to equipment during power dispatching, such as transformer temperature and line aging. Equipment operating status can be collected through IoT monitoring nodes.

[0032] It should be noted that initial power data includes, but is not limited to, the examples above. Other data signals associated with power dispatching can also be used as initial power data.

[0033] In one specific embodiment of step S101, the aforementioned initial power data is continuously collected and uploaded to the central processing unit through various sensors and automation systems.

[0034] Step S102: Perform data fusion on the initial power data to construct a power data matrix.

[0035] Step S102 is a crucial step in the deep processing and integration of the initial power data to form a power data matrix. This step mainly involves unifying the representation of various discrete or independent heterogeneous data into a multi-dimensional matrix structure, namely the power data matrix.

[0036] Specifically, initial power data can include data and corresponding timestamps. Since the initial power data includes various data types, with different sampling time intervals and data types, these data need to be processed and fused to obtain a power data matrix before it can be used for subsequent model analysis.

[0037] In a specific embodiment of step S102, a set of attribute values ​​recorded daily and hourly, including different types such as electricity consumption (MW), temperature readings (°C), and wind speed levels, can be converted and normalized to a specific range to construct a time-series related vector set, which then becomes the row elements of a matrix. Such a matrix not only includes historical trends but also reflects the current overall operating condition of the power grid in real time, facilitating subsequent in-depth calculations.

[0038] Step S103: Input the power data matrix into the risk assessment model to obtain the risk assessment result; wherein, the risk assessment model is used to extract risk features from the power data matrix from multiple dimensions, and obtain the risk assessment result based on the risk features, and the risk assessment model is also used to dynamically allocate the weights of the risk features.

[0039] Among them, risk characteristics are features extracted from the content of the power data matrix that are related to the risk assessment results, such as short-term load fluctuation characteristics, meteorological anomaly correlation characteristics, and equipment health characteristics.

[0040] Risk assessment results can be used to indicate whether there are risks in power dispatching in the current or future period.

[0041] In one specific embodiment, the risk assessment results may include risk level and / or risk management strategy.

[0042] For example, risk levels can include high risk, medium risk, and low risk; risk types can include power transmission and distribution bottlenecks caused by abnormal peak electricity consumption, and power shortages caused by a sudden decrease in photovoltaic efficiency due to rainstorms or thunderstorms; risk management strategies include activating backup power supplies if there is a risk of power shortages.

[0043] In step S103, a pre-trained risk assessment model with high adaptability and self-correction capabilities is used to predict the risks of power dispatch. This model, based on training logic, can dynamically extract features from the continuously input power data matrix and calculate a total score representing the risk assessment result by dynamically adjusting the weights of different risk features.

[0044] pass Figure 1This paper proposes a power dispatch risk early warning method based on fused power data. It acquires multiple initial power data sets and performs fusion analysis on these data to conduct a more comprehensive risk analysis of the power data. Based on dynamic risk assessment, it provides early warnings of power grid operation risks and adjusts the power grid's operating status accordingly. This scheme can also adjust the model's weights for each risk feature in real time based on changes in the power data matrix, making the risk assessment results more adaptable to the complex changes in power data and avoiding the impact of coupling relationships between related data on the accuracy of the risk assessment results. Therefore, it can comprehensively consider multiple influencing factors and improve the accuracy of power dispatch risk management.

[0045] In one specific embodiment, the risk assessment model can be built on a Long Short Term Memory (LSTM) network architecture. This is because information in the power dispatching system may have long-term correlations, and LSTM can handle long-term dependencies well, making it more suitable for risk analysis scenarios in power dispatching systems.

[0046] Compared to conventional LSTM models, the risk assessment model in this application can not only dynamically extract risk features from the power data matrix, but also dynamically adjust the weights of different risk features in real time based on the correlation between risk features and the online training of the model. This enables accurate analysis of power data matrices containing complex data, improving the accuracy of power dispatch risk early warning based on fused power data.

[0047] In an optional embodiment, please refer to Figure 2 The risk assessment model can dynamically allocate the weights of the risk characteristics through the following steps:

[0048] Step S201: Standardize the power data matrix according to a sliding time window.

[0049] Specifically, a time decay factor can be used to standardize the power data matrix within a sliding time window. For example, if the sliding time window is 30 minutes, the standardization parameters can be calculated using only the power data matrix within the most recent time window. Time decay weights are assigned to the data within this window; newer data has a greater impact on the mean (μ) and standard deviation (σ).

[0050] Furthermore, the window length can be automatically adjusted based on data volatility. For example, when a sudden increase in variance is detected (such as lightning activity), the sliding time window is shortened from 30 minutes to 10 minutes.

[0051] Step S201 achieves the following effects: Real-time sensitivity; that is, through exponential decay, standardized parameters can quickly respond to the latest data changes (e.g., updating thresholds within 5 seconds after a lightning strike). Resistance to historical interference; that is, reducing the weight of old data and preventing historically stable data from masking current risks. Interpretability. The weight allocation conforms to the physical intuition of "nearer is larger, farther is smaller," making it easy for operations and maintenance personnel to understand.

[0052] Step S202: Calculate the Mahalanobis distance between different risk characteristics based on the standardized power data matrix.

[0053] Step S203: Dynamically allocate the weights of the risk features based on the Mahalanobis distance.

[0054] Because power data features have large differences in dimensions (e.g., voltage is in kV, wind speed is in m / s, and temperature is in °C), Euclidean distance can lead to calculation results being dominated by large-scale features due to the different dimensions. Mahalanobis distance, by standardizing the data through the covariance matrix, eliminates the difference in dimensions. By replacing Euclidean distance with Mahalanobis distance, it can solve the problem of weight distortion caused by feature correlation, thereby resolving the complex correlation between power data features and significantly improving the accuracy of risk warning.

[0055] Furthermore, the covariance matrix can be updated with sliding window data at preset intervals (e.g., 5 minutes) to update the Mahalanobis distance, thereby capturing real-time changes in the characteristic relationship. In a specific example, when the photovoltaic output in a certain area suddenly increases, the correlation between load and light intensity strengthens. After dynamically adjusting the Mahalanobis distance, normal fluctuations can be avoided from being misjudged as risks.

[0056] In another alternative embodiment, please refer to Figure 3 The risk assessment model can also dynamically allocate the weights of the risk characteristics through the following steps:

[0057] Step S301: Determine the core variables that affect weight allocation.

[0058] The goal of this step is to identify the variables most decisive for early warning outcomes from all factors related to power dispatch risk, such as grid load, weather changes, and historical fault data. The selection of these core variables is based on expert experience and correlation analysis of statistical data. Only by clearly identifying these key variables can subsequent steps establish an effective dynamic adjustment mechanism around them.

[0059] Step S302: Track the data fluctuations of the core variables in real time to obtain the change patterns.

[0060] This step involves real-time tracking of data fluctuations for each variable and recording points of abnormal change. During this process, a real-time monitoring system in computer equipment continuously monitors the numerical changes of core variables. Once a variable's fluctuation exceeds a predefined range or exhibits behavior outside of a predictable pattern, it is marked as an anomaly, providing data support for subsequent weight readjustment. For example, for the variable of grid load, if a sudden peak fluctuation exceeds 150% of the normal level, it must be identified as a key anomaly to be monitored.

[0061] Step S303: Set the update frequency of the weight allocation according to the change pattern.

[0062] This step involves resetting the update frequency of the weights based on the changing patterns of the core variables. By statistically analyzing the fluctuations of these core variables, the length of their update cycle T is determined. This value dynamically changes with the stability of the variables: when the core variables fluctuate significantly, T is shortened for timely response; conversely, if the core variables tend to stabilize, the interval of T is lengthened. The update cycle threshold C is a fixed value, typically set to 24 hours in practice, but may be adjusted according to requirements. The optimal value of the C parameter needs to be determined by fitting historical data for the specific scenario.

[0063] In a specific embodiment of step S303, if the update frequency T≥C, then the weight update is calculated based on the variable importance index F; where T represents the update cycle length, C represents the update cycle threshold parameter, and F is a value obtained based on the historical fluctuation statistics of the variable, and the formula is (update condition is met when T≥C).

[0064] This embodiment calculates weight updates based on the variable importance index F, which occurs when the update condition is met (i.e., when the update period T ≥ C). The core of the formula is to ensure that the weight adjustment of each core variable is related to its actual performance. The F value is obtained from the historical fluctuation statistics of the variable, reflecting the degree of influence of a core variable on the overall system within a past time window. For example, in one embodiment, if the grid load fluctuates drastically and frequently, its F value will be correspondingly larger, leading to a significant increase in its assigned weight.

[0065] Here's a specific example: For instance, a city uses this method for power dispatch risk early warning based on fused power data. The core variables are temperature and load power, which are collected and stored by the meteorological forecasting system and power distribution equipment, respectively. Assume that temperature fluctuations are small, but load power suddenly increases. Records show that the temperature fluctuation rate remains at 3%, while the load increases more than fivefold. At this point, T = 8 hours is less than the default C = 24 hours, and the update condition is not met. However, an update will be forced after two consecutive abnormal load fluctuations. Based on the statistically derived high load F, a higher weight is allocated to this key variable to optimize the dispatch strategy. The formula is reasonably designed because it ensures dynamism while considering the importance of variables, allowing the model to always adapt to new situations.

[0066] In an optional embodiment, the risk assessment model described above can be constructed based on random forest and Bayesian network, wherein the random forest is used to process high-dimensional data in the power data matrix to identify the correlation between the nonlinear risk features, and the Bayesian network is used to update the risk assessment results in conjunction with the real-time power data matrix.

[0067] First, the data input sources and data preprocessing methods for the random forest module in the risk assessment model are determined. This step aims to ensure that the high-dimensional power data matrix meets the requirements of the random forest algorithm to effectively identify potential nonlinear correlations. Random forests use multiple trees to classify or regress data, thereby reducing the impact of noise on the final results. For example, in practical applications, a high-dimensional data matrix formed by collecting historical load, temperature fluctuations, line capacity, and other factors from the power dispatching system is used, and outliers are removed to ensure interpretability and consistency.

[0068] Next, a random forest model is constructed and trained to process high-dimensional power data matrices, extracting the complex relationships between nonlinear risk features. This process uses randomly extracted features to ensure the diversity of each decision tree, and finally, the results from multiple trees are combined to output a stable risk assessment probability value. The number of trees in the random forest is an important hyperparameter. Let this parameter be T, typically ranging from 10 to 500. After multiple validations, its optimal setting is T = 200, at which point the model achieves a balance between complexity and accuracy.

[0069] Next, a Bayesian network is designed and built to update the risk assessment results initially generated by the random forest. This step utilizes a probabilistic model to capture conditional dependencies. For example, by inputting real-time data matrices, such as the aging degree of transmission equipment, as new conditions into the established nodes, the joint probability distribution under each condition is readjusted, thereby completing the dynamic calibration.

[0070] Finally, after integrating the results of the two parts mentioned above, the system was deployed in practical testing scenarios to verify its effectiveness. For example, in one embodiment, when facing an upcoming peak load period, the random forest quickly identifies specific time periods where insufficient power generation is likely to cause local faults, and continuously revises the predicted probability of risk occurrence using Bayesian methods with the help of real-time weather warning data. This combination of technologies enables the formulation of more intelligent and accurate power dispatch management strategies.

[0071] Furthermore, the parameters of the risk assessment model can be optimized through transfer learning. This allows for the optimization of model parameters using historical failure data, improving generalization ability in small sample scenarios.

[0072] In an optional embodiment, the data in the power data matrix may include the following dimensions: time dimension, spatial dimension, and feature channel.

[0073] Because the spatial granularity of the initial power data collected is inconsistent (e.g., meteorological data is a regional grid, while power grid equipment data is specific coordinate points), it is necessary to establish a mapping relationship between electrical distance and geographical location to unify the spatial dimension of the data in the power data matrix.

[0074] Specifically, the spatial dimension of the data in the power data matrix can be generated based on the following steps: dividing the power grid coverage area into a dynamic spatial grid; mapping the initial power data to the dynamic spatial grid based on spatial association rules; and spatially fusing the initial power data.

[0075] Dynamic spatial grid division based on power grid coverage area refers to dividing virtual grid units according to specific rules within the geographical area covered by the actual power grid. The characteristic of dynamic spatial grids is that their size and shape can be flexibly adjusted based on load demand, terrain factors, or seasonal load characteristics. Dynamic spatial grids control the granularity of grid division through parameters; in the optimal case, fixed parameters can be set to ensure denser segmentation in high-risk areas.

[0076] Mapping the initial power data to the dynamic spatial grid based on spatial association rules is another crucial step. Initial power data, such as current, voltage, and load, are mapped to each of the generated grid cells according to preset logical rules. These spatial association rules may be based on nearest neighbor methods or other complex algorithms. This process reduces distortion caused by directly using single-point locations and improves overall accuracy. In one embodiment, when analyzing the impact of summer high temperatures on power supply stability, the initial power parameters collected from all substations are aggregated into their respective dynamic grids, and combined with meteorological data to achieve more accurate data matching.

[0077] Finally, taking risk warning in power dispatch as an example, specifically, when faced with a sudden increase in local load caused by sudden weather changes, if the detection results of a certain dynamic grid show that abnormal values ​​occur in a concentrated manner (i.e., exceeding the upper limit of the set threshold), a warning message is issued in a timely manner to assist in making emergency control decisions.

[0078] The initial power data collected suffers from the following issues: significant differences in collection frequency and latency between different data sources (e.g., grid data is in the millisecond range, meteorological data in the hourly range, and equipment status data in the minute range); and inconsistent data timestamps (e.g., delayed meteorological data release, and asynchronous equipment clocks). Therefore, the initial power data needs to be processed as follows to ensure that the time dimension of the data in the generated power data matrix is ​​aligned. Please refer to [link to relevant documentation]. Figure 4 That is, the time dimension of the data in the power data matrix can be generated based on the following steps:

[0079] Step S401: Interpolate the multiple initial power data according to a sliding window to generate multiple consecutive intermediate data.

[0080] Specifically, a unified reference time axis can be set (such as using the UTC time of the power grid dispatching system as a benchmark), and a sliding window interpolation method can be used for asynchronous data. For example, for hourly updated meteorological data, linear interpolation or cubic spline interpolation can be used within the window (such as 5 minutes before and after) to align with the second-level timestamp of the power grid data.

[0081] Step S402: Downsample the high-frequency intermediate data.

[0082] For example, millisecond-level power grid data can be aggregated by minute-level averages and matched with equipment status data.

[0083] Step S403: Upsample the low-frequency intermediate data.

[0084] For example, meteorological data can be interpolated to minute-level values ​​using time series forecasting models (such as LSTM).

[0085] Steps S402 and S403 align the sampling times of high-frequency and low-frequency sampled data. During steps S402 and S403, Dynamic Time Warping (DTW) can be used to align sequences with different sampling rates (such as user load curves and meteorological change curves). An adaptive sliding window can also be set to dynamically adjust the window size based on data characteristics (e.g., reducing the window to 1 minute during thunderstorms).

[0086] Step S404: Complete the missing values ​​of the intermediate data.

[0087] Specifically, missing data caused by communication interruptions can be filled in. For example, for short-term missing data (such as data missing due to communication interruption time < 5 minutes), spatiotemporal kriging interpolation can be used for filling (combined with data from neighboring sites). For long-term missing data (such as data missing due to communication interruption time greater than or equal to 5 minutes), it can be marked as an anomaly and manual intervention can be triggered.

[0088] It should be noted that there is no sequential relationship between steps S402, S403, and S404. Steps S402 and S403 can both be executed, or only one of them can be executed. The specific execution can be set according to the type of initial power data and the sampling frequency.

[0089] This method acquires and integrates various initial power data, adjusting the power grid's operating status based on dynamic risk assessment. This approach effectively addresses issues such as power supply stability due to excessive peak-to-valley differences, power imbalance caused by intermittent renewable energy generation, the risk of critical equipment outages due to overload, the social impact of uneven load shedding, and security protection of the dispatch system against cyberattacks.

[0090] The specific embodiments described above further illustrate the purpose, technical solutions, and beneficial effects of the embodiments of this disclosure. It should be understood that the above descriptions are merely specific embodiments of the embodiments of this disclosure and are not intended to limit the scope of protection of the embodiments of this disclosure. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the embodiments of this disclosure should be included within the scope of protection of the embodiments of this disclosure.

[0091] In one embodiment, please refer to Figure 5 This application also provides a power dispatch risk early warning device 50 based on fused power data, comprising:

[0092] The data acquisition module 501 is used to acquire initial power data, which includes at least power grid operation data, user electricity consumption data, meteorological data and equipment status data;

[0093] Data fusion module 502 is used to fuse the initial power data and construct a power data matrix;

[0094] The evaluation module 503 is used to input the power data matrix into the risk evaluation model to obtain the risk evaluation result; wherein, the risk evaluation model is used to extract risk features from the power data matrix from multiple dimensions and obtain the risk evaluation result based on the risk features, and the risk evaluation model is also used to dynamically allocate the weights of the risk features.

[0095] In an optional embodiment, the power dispatch risk early warning device 50 based on fused power data further includes:

[0096] A standardization module is used to standardize the power data matrix according to a sliding time window;

[0097] The Mahalanobis distance calculation module is used to calculate the Mahalanobis distance between different risk characteristics based on the standardized power data matrix.

[0098] The weight allocation module is used to dynamically allocate the weights of the risk features based on the Mahalanobis distance.

[0099] In an optional embodiment, the power dispatch risk early warning device 50 based on fused power data further includes:

[0100] The core variable determination module is used to determine the core variables that affect weight allocation;

[0101] The tracking module is used to track the data fluctuations of the core variables in real time and obtain the patterns of change;

[0102] The weight update module is used to set the update frequency of the weight allocation according to the change pattern.

[0103] In an optional embodiment, the risk assessment model is constructed based on a random forest and a Bayesian network, wherein the random forest is used to process high-dimensional data in the power data matrix to identify the correlations between the nonlinear risk features, and the Bayesian network is used to update the risk assessment results in conjunction with the real-time power data matrix.

[0104] In an optional embodiment, the parameters of the risk assessment model are optimized using a transfer learning mechanism.

[0105] In an optional embodiment, the data in the power data matrix includes the following dimensions: time dimension, spatial dimension, and feature channel.

[0106] In an optional embodiment, the data fusion module 502 includes:

[0107] Grid partitioning unit, used to divide dynamic spatial grids according to the power grid coverage area;

[0108] The spatial fusion unit is used to map the initial power data to the dynamic spatial grid based on spatial association rules, and to perform spatial fusion on the initial power data.

[0109] In an optional embodiment, the data fusion module 502 further includes:

[0110] An interpolation unit is used to interpolate multiple initial power data according to a sliding window to generate multiple consecutive intermediate data.

[0111] A sampling alignment unit is used to downsample the high-frequency intermediate data and / or upsample the low-frequency intermediate data.

[0112] The missing value completion unit is used to complete the missing values ​​of the intermediate data.

[0113] For more information on the working principle and operation mode of the power dispatch risk early warning device 50 based on integrated power data, please refer to the relevant information. Figures 1 to 4 The relevant description of the power dispatch risk early warning method based on fused power data will not be repeated here.

[0114] This application also provides a storage medium, specifically a computer-readable storage medium, for storing a computer program, wherein the program executes the steps of any of the above-described power dispatch risk early warning methods based on fused power data by a computer or processor. The computer-readable storage medium may include non-volatile or non-transitory memory, and may also include optical discs, hard disk drives, solid-state drives, etc.

[0115] This application also provides a computer device, which may include a memory and a processor. The memory stores a computer program, and when the program is executed by the processor, it can implement any of the steps of the power dispatch risk early warning method based on fused power data described above.

[0116] In the embodiments of this application, the processor may be a general-purpose processor, a digital signal processor, an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components, capable of implementing or executing the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor may be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this application can be directly manifested as being executed by a hardware processor, or executed by a combination of hardware and software modules within the processor.

[0117] In the embodiments of this application, the memory can be non-volatile memory, such as a hard disk drive (HDD) or a solid-state drive (SSD), or it can be volatile memory, such as random-access memory (RAM). Memory is any other medium capable of carrying or storing desired program code in the form of instructions or data structures, and accessible by a computer, but is not limited thereto. The memory in the embodiments of this application can also be a circuit or any other device capable of implementing storage functions for storing computer programs and / or data.

[0118] The power dispatch risk early warning method based on fused power data provided in this application can be implemented entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program is loaded and executed on a computer, it generates all or part of the processes or functions described in the embodiments of this application. The computer can be a general-purpose computer, a special-purpose computer, a computer network, a network device, a user equipment, or other programmable device. The computer program can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that integrates one or more available media. The available media may be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., digital video discs (DVDs)), or semiconductor media (e.g., SSDs), etc.

[0119] The above detailed introduction clearly expresses the operational details and technical framework of the power dispatch risk early warning method based on integrated power data.

[0120] This invention presents a power dispatch risk early warning method based on fused power data, designed to comprehensively improve the stability and security of power dispatch. Specifically, this method primarily acquires initial power data (such as grid operation data, user electricity consumption data, meteorological data, and equipment status data), and then fuses this data to construct a power data matrix. The constructed power data matrix is ​​then input into a risk assessment model to dynamically extract risk features, and different features are assigned corresponding weights based on real-time data analysis results, thereby accurately predicting and assessing potential dispatch risks.

[0121] The above process, with the help of computer equipment and storage media, enables efficient risk prediction and control strategy output, effectively reducing power system operation problems caused by various complex factors and improving the overall intelligence level of dispatching.

[0122] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

[0123] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0124] In the above embodiments, the descriptions of each embodiment have their own emphasis, and any multiple embodiments can be used in combination. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0125] In implementation, each step of the above method can be completed by integrated logic circuits in the processor's hardware or by instructions in software. The steps of the method disclosed in the embodiments of this application can be directly manifested as execution by a hardware processor, or as a combination of hardware and software units within the processor. The software units can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory, and the processor executes the instructions in the memory, combining them with its hardware to complete the steps of the above method. To avoid repetition, detailed descriptions are omitted here.

[0126] In the embodiments of this application, the processor of the above-described device may be a Central Processing Unit (CPU), which may also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc.

[0127] This application also provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program operable to cause a computer to perform some or all of the steps of any of the methods described in the above method embodiments. This computer program product can be a software installation package.

[0128] In the several embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical or other forms.

[0129] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of the embodiments of this application, depending on actual needs.

[0130] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0131] If the aforementioned integrated units are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, 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. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or TRP, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0132] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this article indicates that the preceding and following related objects have an "or" relationship.

[0133] In the embodiments of this application, "multiple" refers to two or more.

[0134] The descriptions of "first," "second," etc., appearing in the embodiments of this application are for illustrative purposes and to distinguish the objects being described. They have no order and do not indicate any special limitation on the number of devices in the embodiments of this application, nor do they constitute any limitation on the embodiments of this application.

[0135] In this application, the term "connection" refers to various connection methods, such as direct connection or indirect connection, to achieve communication between devices. This application does not impose any limitations on this.

[0136] The embodiments of this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A power dispatch risk early warning method based on fused power data, characterized in that, The method includes: Acquire initial power data, which includes at least grid operation data, user electricity consumption data, meteorological data, and equipment status data; The initial power data is fused to construct a power data matrix; The power data matrix is ​​input into the risk assessment model to obtain the risk assessment results; The risk assessment model is used to extract risk features from the power data matrix from multiple dimensions and obtain risk assessment results based on the risk features. The risk assessment model is also used to dynamically allocate the weights of the risk features.

2. The method according to claim 1, characterized in that, The risk assessment model dynamically allocates the weights of the risk characteristics through the following steps: The power data matrix is ​​standardized according to a sliding time window; The Mahalanobis distance between different risk characteristics is calculated based on the standardized power data matrix. The weights of the risk features are dynamically allocated based on the Mahalanobis distance.

3. The method according to claim 1, characterized in that, The risk assessment model dynamically allocates the weights of the risk characteristics through the following steps: Identify the core variables that influence weight allocation; Real-time tracking of the data fluctuations of the core variables reveals patterns of change. The update frequency of the weight allocation is set according to the aforementioned change pattern.

4. The method according to any one of claims 1 to 3, characterized in that, The risk assessment model is built based on random forest and Bayesian network, wherein the random forest is used to process high-dimensional data in the power data matrix to identify the correlation between the nonlinear risk features, and the Bayesian network is used to update the risk assessment results in combination with the real-time power data matrix.

5. The method according to claim 4, characterized in that, The parameters of the risk assessment model are optimized through transfer learning.

6. The method according to claim 1, characterized in that, The data in the power data matrix includes the following dimensions: time dimension, spatial dimension, and feature channel.

7. The method according to claim 6, characterized in that, The spatial dimension of the data in the power data matrix is ​​generated based on the following steps: Dynamic spatial grids are divided based on the power grid coverage area; The initial power data is mapped to the dynamic spatial grid based on spatial association rules, and the initial power data is spatially fused.

8. The method according to claim 6 or 7, characterized in that, The time dimension of the data in the power data matrix is ​​generated based on the following steps: Multiple initial power data are interpolated using a sliding window to generate multiple consecutive intermediate data. The intermediate data at high frequencies is downsampled, and / or the intermediate data at low frequencies is upsampled; Complete the missing values ​​in the intermediate data.

9. A computer device comprising a processor and a storage device, said storage device being adapted to store a plurality of program codes, characterized in that, The program code is adapted to be loaded and run by the processor to perform the method of any one of claims 1 to 8.

10. A storage medium storing a plurality of program codes, characterized in that, The program code is adapted to be loaded and run by a processor to perform the method of any one of claims 1 to 8.

Images