Power supply equipment warning method and device and computer program product
By integrating multi-source data and CNN-LSTM models, the accuracy and real-time performance of alarm mechanisms for power supply equipment were addressed, enabling comprehensive, real-time monitoring and precise alarms for the equipment, thereby reducing the risk of equipment failure and power outages.
Patent Information
- Application Number
- CN202511012764.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-22
- Publication Date
- 2025-11-07
AI Technical Summary
Existing alarm mechanisms for power supply equipment are insufficient to effectively address external environmental risks, especially in situations such as high temperatures, severe weather, and floods. They cannot provide comprehensive, real-time, and accurate alarms, leading to equipment failures and power supply interruptions.
By integrating meteorological, water level, internal equipment video, and external video data, a CNN-LSTM hybrid model is used to extract water accumulation features and perform weighted calculations to generate the equipment's alarm level.
It enables comprehensive, real-time monitoring of power supply equipment, significantly improving alarm accuracy, reducing false alarm rate, minimizing manual intervention, shortening response time, and ensuring stable operation of the power system.
Smart Images
Figure CN120913356A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of power systems, in particular to a power supply guaranteeing device alarm method, device and computer program product. BACKGROUND
[0002] In the power system, the stable operation of the power supply guaranteeing device is the core prerequisite for ensuring reliable power supply, and its operation state directly determines the support ability of the power system to social production and life.
[0003] As a key period of power supply, the power supply guaranteeing work faces multiple severe challenges in summer: on the one hand, the continuous high temperature weather leads to a substantial increase in electricity load, testing the operation limit of the equipment; on the other hand, severe weather conditions such as heavy rain, strong wind, lightning and other adverse weather conditions occur frequently, and the abnormal rise of water level in the flood season and other hydrological changes jointly pose a serious threat to the power infrastructure. Specifically, key equipment such as substation facilities, transmission line towers, distribution rooms and other key equipment are prone to failure due to problems such as rainwater accumulation, wind tower collapse, lightning breakdown, flood immersion, and even cause regional power supply interruption, seriously affecting the social production and life order.
[0004] However, the current power supply guaranteeing device alarm mechanism still has obvious limitations and is difficult to effectively respond to the above external environmental risks. The traditional alarm mode mainly relies on two types of data to determine: one is the operation parameters of the equipment itself (such as temperature, current, voltage, etc.), and the other is the single weather index of the location where the equipment is located (such as rainfall, wind speed, etc.). This mode can only achieve simple threshold judgment at the numerical level, and lacks accurate perception and confirmation of the real environment state around the equipment - for example, it is difficult to scientifically estimate the actual water depth and spreading trend around the equipment through rainfall data, and it is difficult to predict the flooding risk of low-lying area equipment.
[0005] More prominent is that the existing water level monitoring means has significant technical shortcomings: the water level sensor only has a local alarm function and is not linked with the power grid topology and dispatching system. When key nodes such as low-lying substations are threatened by floods, the system cannot automatically trigger emergency responses such as dynamic load adjustment and drainage equipment start-up in combination with real-time water level data, and still needs to rely on manual patrol and on-site disposal, which is easy to expand the scope of failure due to delayed response. SUMMARY
[0006] The technical problem to be solved by the embodiments of the present application is to provide a power supply guaranteeing device alarm method, device and computer program product to realize all-round and real-time monitoring and accurate alarm of the power supply guaranteeing device.
[0007] To solve the above technical problems, the present application provides a power supply guaranteeing device alarm method, comprising:
[0008] acquire and process multi-source data related to the power supply equipment, the multi-source data at least including meteorological data, water level monitoring data, internal video data of the equipment and external video data of the equipment, and convert the meteorological data and the water level monitoring data into corresponding quantitative indicators;
[0009] analyze the internal video data of the equipment and the external video data of the equipment, extract water accumulation related features, and obtain corresponding video risk scores through a preset model processing;
[0010] According to a preset weight rule, the meteorological data, the water level monitoring data corresponding to the quantitative indicators and the video risk scores are weighted and calculated to obtain the final alarm score of the power supply equipment, and the corresponding alarm level is output.
[0011] Preferably, the meteorological data is grid meteorological data covering the area where the power supply equipment is located, and the accuracy of the grid meteorological data is 1km×1km.
[0012] Preferably, the process of converting the meteorological data into corresponding quantitative indicators includes: matching the latitude and longitude data of the power supply equipment with the grid meteorological data to obtain the rainfall data of the area where the equipment is located, and converting the rainfall data into a quantitative indicator of 0-100.
[0013] Preferably, the water level monitoring data is the real-time water level data of the monitoring station with the highest water level in a fixed range around the power supply equipment, and the process of converting the water level monitoring data into corresponding quantitative indicators includes: calculating the difference between the real-time water level and the warning water level, and converting the difference into a quantitative indicator of 0-100.
[0014] Preferably, before analyzing the internal video data of the equipment and the external video data of the equipment, the video data is preprocessed to remove noise and correct light.
[0015] Preferably, the water accumulation related features include the proportion of the water accumulation area to the region of interest, the relative value of the water accumulation depth, the water accumulation spread speed and the water accumulation flowability.
[0016] Preferably, the preset model is a CNN-LSTM hybrid model, the CNN part of the CNN-LSTM hybrid model adopts a ResNet50 network structure, which is used to extract spatial features of video frames and output feature vectors; the LSTM part is used to receive continuous multiple frames of feature vectors and combine time sequence features to output video risk scores.
[0017] Preferably, the alarm level at least includes three levels of low risk, medium risk and high risk, and each level corresponds to a preset alarm score range.
[0018] The application further provides a power supply equipment alarm device, comprising:
[0019] one or more processors;
[0020] a memory;
[0021] one or more application programs, wherein the one or more application programs are stored in the memory and configured to be executed by the one or more processors, and the one or more application programs are configured to execute the power supply equipment alarm method.
[0022] The application further provides a computer program product comprising computer instructions instructing a computer device to execute the operations corresponding to the method.
[0023] The application has the following beneficial effects: The application brings multiple significant beneficial effects by fusing four types of data, i.e., weather, water level, internal video and external video. In terms of alarm accuracy, compared with the traditional method which only relies on device self parameters or single weather data, the application describes the risks such as water immersion faced by the device from multiple dimensions. The grid weather data accurately matches the device location, the water level data reflects the real hydrological conditions in the surrounding area, the internal and external videos extract the water accumulation features through the CNN-LSTM model, and the multiple source information is verified with each other, so that the alarm research and judgment are closer to the actual environment, the alarm accuracy is greatly improved, and the misjudgment rate is significantly reduced. In terms of monitoring efficiency, real-time and automatic monitoring is realized. The collection of grid weather, water level and video data is automatic, after quantitative conversion, feature extraction and other processing, automatic analysis and scoring are realized through the model and weighted fusion, the whole process does not require manual intervention, the response lag problem caused by the traditional dependence on manual patrol and judgment is changed, and the manual judgment cost is greatly reduced. In practical application, the comprehensive alarm level is directly pushed to the relevant personnel, which provides accurate basis for device disaster prevention and mitigation measures such as starting the drainage device, adjusting the load, personnel evacuation and the like, effectively shortens the response time, effectively reduces the risk of device failure and power interruption, and guarantees the stable operation of the power system. BRIEF DESCRIPTION OF DRAWINGS
[0024] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description only constitute some embodiments of the application, and for those skilled in the art, other drawings can also be obtained from these drawings without creative labor.
[0025] Figure 1 is a flowchart of a power supply equipment alarm method according to an embodiment of the application.
[0026] Figure 2is a specific flowchart of a power supply protection equipment alarm method according to an embodiment of the present application. DETAILED DESCRIPTION
[0027] The following description of the embodiments is with reference to the drawings, which illustrate specific embodiments in which the present application can be implemented.
[0028] Referring to Figure 1 The embodiment one of the present application provides a power supply protection equipment alarm method, which comprises the following steps.
[0029] In step S1, multi-source data related to the power supply protection equipment is acquired and processed, wherein the multi-source data at least comprises meteorological data, water level monitoring data, internal video data of the equipment and external video data of the equipment, and the meteorological data and the water level monitoring data are converted into corresponding quantitative indexes.
[0030] In step S2, the internal video data of the equipment and the external video data of the equipment are analyzed, water accumulation related features are extracted, and corresponding video risk scores are obtained through preset model processing.
[0031] In step S3, according to a preset weight rule, the quantitative indexes corresponding to the meteorological data and the water level monitoring data and the video risk scores are weighted and calculated to obtain a final alarm score of the power supply protection equipment, and a corresponding alarm level is output.
[0032] Specifically, the embodiment of the present application innovatively integrates meteorological monitoring data and surrounding environmental data, including equipment, rainfall data, and water level monitoring data, internal video data of the power room and surrounding video data of the power room, improves the accuracy of the power supply protection equipment alarm, reduces false alarms, and the research and judgment of the alarm is closer to the real environment, reduces the work of human judgment, and realizes the all-round and real-time monitoring and accurate alarm of the power supply protection equipment.
[0033] Please combine Figure 2As shown, in the specific implementation process of the present application, step S1 first carries out the acquisition and processing of multi-source data. For meteorological data, the system acquires 1km*1km grid point meteorological data covering the set range (for example, a certain city), and at the same time calls the account information of the power supply equipment, from which the latitude and longitude data of the equipment are extracted. Through the preset coordinate conversion rule, such as the conversion algorithm based on Gauss-Kruger projection, the latitude and longitude data are converted into X-axis and Y-axis coordinates matching the grid point meteorological data, the real-time rainfall data of the area where the equipment is located is located in the grid point meteorological data according to the converted coordinates, and the rainfall value is converted into a quantitative index of 0-100 according to the set rule, which reflects the potential risk degree of the impact of rainfall on the equipment. The rule of converting rainfall into a quantitative index of 0-100 here can be set according to the correlation between rainfall and equipment waterlogging risk in historical data, for example, when the rainfall reaches a certain threshold, the quantitative index increases linearly with the increase of rainfall, so as to intuitively reflect the potential risk degree of the impact of rainfall on the equipment. For water level monitoring data, the highest monitoring station within a fixed range (such as a circular area with the equipment as the center and a radius of 500 meters) around the power supply equipment is selected as the data source, the real-time water level data of the monitoring station is collected, the difference between the water level and the warning water level of the area is calculated, and the water level difference is converted into a quantitative index of 0-100 through a preset rule, the quantitative index is 0 when the water level difference is 0, and the quantitative index tends to 100 when the water level approaches the warning water level, so as to clearly reflect the water level threat degree of the equipment.
[0034] Meanwhile, the internal video data and the external video data of the power supply equipment are processed and analyzed by step S2 to obtain a more actual waterlogging situation. For the internal video data (such as the video in the power room) and the external video data (such as the video around the power room), first, a preprocessing operation is performed, including noise removal processing to eliminate interference information in the image, and light correction to ensure the accuracy of image analysis under different light conditions. The noise removal processing can use a median filter algorithm to eliminate interference information such as salt and pepper noise in the image, and the light correction can use a histogram equalization method to ensure that the image brightness and contrast remain consistent under different light conditions, thereby ensuring the accuracy of image analysis. In the preprocessed video image, an image segmentation algorithm is used to automatically identify and retain areas that may appear waterlogging (such as the ground, corners, etc.), while ignoring areas unrelated to waterlogging monitoring (such as equipment in the room, signal lights, sky, trees, etc.). It can be understood that the image segmentation algorithm can be based on a semantic segmentation model of deep learning, and can achieve accurate identification after training on a large number of sample images containing waterlogging areas. From the selected effective areas, waterlogging-related direct features are extracted, including the proportion of the waterlogging area to the region of interest (ROI), the relative value of the waterlogging depth (which can be calculated by comparing the known height object in the image with the waterlogging), the waterlogging spread speed (which can be calculated by the change of the waterlogging area in adjacent frames of images in a unit of time), and the water flow of the waterlogging (which can be determined by analyzing the texture change of the water surface in the image to determine whether it is dynamic waterlogging).
[0035] Subsequently, the extracted features are input into a preset CNN-LSTM hybrid model for processing, which can effectively combine spatial and temporal features to analyze waterlogging risks. The CNN part uses a ResNet50 network structure to extract spatial features from each frame of image in the video, obtaining a 128-dimensional feature vector that contains texture, edge, and other detailed information of the waterlogging area, such as the reflective texture of the water surface, the edge profile of the water and the ground, etc. The LSTM part receives the continuous 12-frame feature vectors output by the CNN, and combines structured time series features such as waterlogging spread speed to learn the change rule of waterlogging state in the time dimension by means of its gating mechanism, for example, whether the waterlogging is in a slow spread or a rapid spread state. By processing the video data for 5 minutes through the hybrid model, the video risk score output by the model is also between 0 and 100, and the waterlogging risk assessment results of the internal and external equipment are obtained, and the higher the score, the greater the waterlogging risk.
[0036] After the above data processing and analysis is completed, enter the alarm score calculation and result output phase of step S3, and realize effective fusion of multi-source data through weighted calculation. According to the alarm rule feature weight preset according to the actual application scene and historical data, the weights of meteorological data, water level monitoring data, device internal video risk score and device external video risk score can be respectively set as 20%, 30%, 25% and 25%. The specific weight can be adjusted according to the importance of different factors to the device failure. According to the weight, the quantified indexes corresponding to the meteorological data, the quantified indexes corresponding to the water level monitoring data, the device internal video risk score and the device external video risk score are weighted calculated, and the final alarm score of the power supply protection device is obtained. Based on the final alarm score, the alarm level is divided into multiple levels, for example, 0-30 points is low risk, 31-60 points is medium risk, and 61-100 points is high risk. The analysis result containing the alarm level is intelligently pushed to the terminal equipment of the power supply protection personnel and the power supply protection command and dispatch personnel through the internal communication system of the power grid, which provides precise data support for the disaster prevention and reduction protection work of the power supply protection device, and helps rapid response and disposal.
[0037] Corresponding to the power supply protection device alarm method of the foregoing embodiment one of the application, the embodiment two of the application further provides a power supply protection device alarm device, comprising:
[0038] one or more processors;
[0039] a memory;
[0040] one or more application programs, wherein the one or more application programs are stored in the memory and configured to be executed by the one or more processors, and the one or more application programs are configured to execute the power supply protection device alarm method of the foregoing embodiment one of the application.
[0041] Corresponding to the power supply protection device alarm method of the foregoing embodiment one of the application, the embodiment three of the application further provides a computer program product, comprising computer instructions, the computer instructions instruct a computer device to execute the operation corresponding to the power supply protection device alarm method of the foregoing embodiment one of the application.
[0042] Preferably, the processor can be a central processing unit (CPU), and can also be other general-purpose processors, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, etc. The general-purpose processor can be a microprocessor, or the processor can also be any conventional processor. The processor is a control center of the device, and connects various parts of the device through various interfaces and lines.
[0043] The memory mainly includes a program storage area and a data storage area. The program storage area can store an operating system, at least one application required by a function, etc., and the data storage area can store relevant data, etc. In addition, the memory can be a high-speed random access memory, and can also be a non-volatile memory such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc., or the memory can also be other volatile solid-state storage devices.
[0044] It should be noted that the above device can include but is not limited to the processor and the memory, and those skilled in the art can understand.
[0045] From the above description, compared with the prior art, the beneficial effects of the present application are that the present application brings about significant beneficial effects in multiple aspects by fusing four types of data of weather, water level, internal video and external video. In terms of alarm accuracy, compared with the traditional method of relying only on the device's own parameters or single weather data, the present application describes the risks such as water immersion faced by the device from multiple dimensions. The grid point weather data accurately matches the device location, the water level data reflects the real hydrological conditions in the surrounding area, the internal and external videos extract the water accumulation features through the CNN-LSTM model, and the multi-source information is mutually verified, so that the alarm research and judgment are closer to the actual environment, greatly improving the alarm accuracy and significantly reducing the misjudgment rate. In terms of monitoring efficiency, real-time and automated monitoring is realized. The collection of grid point weather, water level and video data is automated, and after quantization conversion, feature extraction and other processing, automatic analysis and scoring are realized through the model and weighted fusion, without manual intervention throughout the process, which changes the response lag problem caused by the traditional reliance on manual patrol and judgment, greatly reduces the cost of manual judgment. In practical application, the comprehensive alarm level is directly pushed to the relevant personnel, providing accurate basis for device disaster prevention and mitigation measures such as starting the drainage device, adjusting the load, personnel evacuation, etc., effectively shortening the response time, effectively reducing the risk of device failure and power interruption, and ensuring the stable operation of the power system.
[0046] The above disclosure is only the preferred embodiment of the present application, and of course cannot limit the scope of the right of the present application, so the equivalent changes made according to the claims of the present application still belong to the scope covered by the present application.
Claims
1. A power supply equipment alarm method characterized by comprising: The method comprises: acquiring and processing multi-source data related to the power supply equipment, the multi-source data at least comprising meteorological data, water level monitoring data, internal video data of the equipment and external video data of the equipment, and converting the meteorological data and the water level monitoring data into corresponding quantitative indicators; analyzing the internal video data of the equipment and the external video data of the equipment, extracting water accumulation related features, and obtaining corresponding video risk scores through a preset model; according to a preset weight rule, performing weighted calculation on the quantitative indicators corresponding to the meteorological data and the water level monitoring data and the video risk scores to obtain a final alarm score of the power supply equipment, and outputting a corresponding alarm level.
2. The method of claim 1, wherein, The meteorological data is grid meteorological data covering the area where the power supply equipment is located, and the accuracy of the grid meteorological data is 1km×1km.
3. The method of claim 2, wherein, The process of converting the meteorological data into corresponding quantitative indicators comprises: matching the latitude and longitude data of the power supply equipment with the grid meteorological data to obtain rainfall data of the area where the equipment is located, and converting the rainfall data into a quantitative indicator of 0-100.
4. The method of claim 1, wherein, The water level monitoring data is the real-time water level data of the monitoring station with the highest water level in a fixed range around the power supply equipment, and the process of converting the water level monitoring data into corresponding quantitative indicators comprises: calculating the difference between the real-time water level and the warning water level, and converting the difference into a quantitative indicator of 0-100.
5. The method of claim 1, wherein, Before analyzing the internal video data of the equipment and the external video data of the equipment, the method further comprises a preprocessing operation of denoising and light correction on the video data.
6. The method of claim 1, wherein, The water accumulation related features include the proportion of the water accumulation area to the region of interest, the relative value of the water accumulation depth, the water accumulation spread speed and the water accumulation flowability.
7. The method of claim 1, wherein, The preset model is a CNN-LSTM hybrid model, the CNN part of the CNN-LSTM hybrid model adopts a ResNet50 network structure, which is used to extract spatial features of video frames and output feature vectors; the LSTM part is used to receive continuous multiple frames of feature vectors and combine time sequence features to output video risk scores.
8. The method of claim 1, wherein, The alarm level at least comprises three levels of low risk, medium risk and high risk, and each level corresponds to a preset alarm score range.
9. A power supply equipment alarm device, characterized in that, The method comprises: one or more processors; a memory; one or more application programs, wherein the one or more application programs are stored in the memory and configured to be executed by the one or more processors, and the one or more application programs are configured to perform the power supply equipment alarm method according to any one of claims 1-8.
10. A computer program product, characterised in that, The computer instructions instruct the computer device to perform operations corresponding to the method according to any one of claims 1-8.