Important power consumer hidden danger troubleshooting and dynamic management and control method and system

Abstract

The invention relates to the technical field of power utilization safety, in particular to an important power consumer hidden danger troubleshooting and dynamic management and control method and system.The method comprises the steps that real-time basic data to be analyzed are obtained, and the real-time basic data to be analyzed comprise electrical equipment operation parameters, environment parameters and user power utilization behavior data; real-time basic data to be analyzed are input to the hidden danger analysis model, a potential safety hazard recognition result is obtained, and the hidden danger analysis model carries out grading early warning based on the potential safety hazard recognition result obtained by training the initial machine learning model through a plurality of samples. According to the method, deep learning is introduced to establish a hidden danger analysis model, electrical equipment operation parameters, environmental parameters and user power consumption behavior data of important power consumers are analyzed, the types and severity of potential safety hazards are accurately identified, the problem that the potential safety hazards of the important power consumers are not checked in place is effectively solved, and the potential safety hazard detection efficiency is improved. The accuracy, timeliness and comprehensiveness of hidden danger checking are remarkably improved, and a powerful guarantee is provided for safe and stable operation of a power system.

Description

Technical Field

[0001] The present invention relates to the technical field of electricity safety, in particular to a method and system for troubleshooting and dynamically controlling hidden dangers of important power users. Background Art

[0002] With the rapid development of my country's economy and power sector, the types and volume of electrical products in use have increased dramatically. This is especially true for key power users (such as hospitals, transportation hubs, and coal and non-coal mines). The safety and stability of electricity use by these users is crucial, impacting every aspect of society. Therefore, real-time and accurate risk detection is essential for these users.

[0003] The current method for troubleshooting hidden dangers for important power users (such as hospitals, transportation hubs, coal mines, and non-coal mines) mainly relies on manual inspections and regular testing. This method has the following problems: (1) Inefficiency: Manual inspections are time-consuming, and data recording and analysis are delayed; (2) Insufficient accuracy: Hidden danger identification relies on the experience of inspectors and is greatly influenced by subjective factors; (3) Management lag: The hidden danger rectification process lacks real-time tracking, making closed-loop management difficult to implement; (4) Lack of visualization: There is a lack of intuitive means to evaluate user hidden dangers, and decision-making lacks data support. Summary of the Invention

[0004] The present invention provides a method and system for detecting and dynamically controlling hidden dangers of important power users, which overcomes the shortcomings of the above-mentioned existing technologies and can effectively solve the problems of low efficiency and failure to detect hidden dangers in a timely manner in existing hidden danger detection methods for important power users.

[0005] One of the technical solutions of the present invention is achieved through the following measures: a method for troubleshooting and dynamically controlling hidden dangers of important power users, comprising: Obtaining real-time basic data to be analyzed, where the real-time basic data to be analyzed includes electrical equipment operating parameters, environmental parameters, and user electricity usage behavior data; Input the real-time basic data to be analyzed into the hidden danger analysis model to obtain the safety hidden danger identification results. The hidden danger analysis model is obtained by training an initial machine learning model using several samples. The initial machine learning model includes a long short-term memory network and a convolutional neural network. Each sample includes electrical equipment operating parameters, environmental parameters, user electricity usage behavior data, and corresponding safety hidden danger type and severity identification information; Provide graded warnings based on safety hazard identification results.

[0006] The following are further optimizations and / or improvements to the above technical solutions: The construction process of the above hidden danger analysis model includes: Obtaining a number of historical hidden danger data, identifying the type and severity of the safety hidden danger corresponding to each historical hidden danger data, obtaining a sample set, and dividing the sample set into a training sample set and a test sample set, wherein the historical hidden danger data includes electrical equipment operating parameters, environmental parameters, and user electricity usage behavior data; The initial machine learning model is trained using the training sample set. Training ends when the training stopping conditions are met to obtain a hidden danger analysis model. The initial machine learning model includes a long short-term memory network, a convolutional neural network, a fully connected layer, and a classifier. The long short-term memory network extracts the long-term dependency features of the input data, and the convolutional neural network extracts the spatial features of the input data. The long-term dependency features and spatial features are fused, and the fused features are classified using a fully connected layer and a classifier to obtain the type and severity of the safety hazard. The trained hidden danger analysis model is tested using the test set to optimize the model parameters of the hidden danger analysis model and output a hidden danger analysis model that meets the test and evaluation requirements.

[0007] The above-mentioned real-time basic data to be analyzed includes: Obtain real-time basic data from important power users, including electrical equipment operating parameters, environmental parameters, and user electricity usage behavior data; Edge computing is introduced to screen and process real-time basic data to establish real-time basic data to be analyzed. Edge computing includes feature extraction through different convolution kernel combinations, data dimensionality reduction using maximum pooling, and integration of convolution results and pooling results using a fully connected layer.

[0008] When performing graded warnings based on the safety hazard identification results, the corresponding graded warning rules include: For general hidden dangers, early warning information will be sent to power users via SMS push, and power users will be reminded to make corrections within the specified time; For major hidden dangers, early warning information will be sent to power users and power management departments through SMS push, APP push, and telephone notifications, and power users will be required to immediately stop the operation of related equipment, make rectifications within the specified time, and take preventive measures.

[0009] The above also includes generating rectification plans based on the safety hazard identification results, including: Identify the type of safety hazard; Extract the corresponding rectification plan from the database according to the type of safety hazard.

[0010] The second technical solution of the present invention is achieved through the following measures: a system for detecting and dynamically controlling hidden dangers of important power users, comprising: A data acquisition unit acquires real-time basic data to be analyzed, wherein the real-time basic data to be analyzed includes electrical equipment operating parameters, environmental parameters, and user electricity consumption behavior data; The hidden danger identification unit inputs the real-time basic data to be analyzed into the hidden danger analysis model to obtain the safety hidden danger identification results. The hidden danger analysis model is obtained by training an initial machine learning model using several samples. The initial machine learning model includes a long short-term memory network and a convolutional neural network. Each sample includes electrical equipment operating parameters, environmental parameters, user electricity usage behavior data, and corresponding safety hidden danger type and severity identification information; The dynamic management and control unit provides graded warnings based on the results of safety hazard identification.

[0011] The following are further optimizations and / or improvements to the above technical solutions: The above-mentioned hidden danger identification unit includes: Model building subunit, including: The sample acquisition module acquires a number of historical hidden danger data, identifies the type and severity of the safety hazard corresponding to each historical hidden danger data, obtains a sample set, and divides the sample set into a training sample set and a test sample set. The historical hidden danger data includes electrical equipment operating parameters, environmental parameters, and user electricity usage behavior data; The model training module uses the training sample set to train the initial machine learning model. When the training stop conditions are met, the training ends and a hidden danger analysis model is obtained. The initial machine learning model includes a long short-term memory network, a convolutional neural network, a fully connected layer, and a classifier. The long short-term memory network extracts the long-term dependency features of the input data, and the convolutional neural network extracts the spatial features of the input data. The long-term dependency features and spatial features are fused, and the fused features are classified using a fully connected layer and a classifier to obtain the type and severity of the safety hazard. The model testing module uses the test set to test the trained hidden danger analysis model, optimizes the model parameters of the hidden danger analysis model, and outputs a hidden danger analysis model that meets the test and evaluation requirements; The hidden danger identification sub-unit inputs the real-time basic data to be analyzed into the hidden danger analysis model to obtain the safety hidden danger identification results.

[0012] The above-mentioned data acquisition unit includes: The acquisition module, including smart meters, load monitoring devices, and various sensors, acquires real-time basic data from important power users, including electrical equipment operating parameters, environmental parameters, and user electricity usage behavior data; The edge computing module extracts features through different combinations of convolution kernels, uses maximum pooling for data dimensionality reduction, and uses a fully connected layer to integrate the convolution results and pooling results to obtain real-time basic data to be analyzed.

[0013] The above-mentioned dynamic control unit includes: A graded warning module provides graded warnings and light warnings based on the results of potential safety hazards identification; The corresponding hierarchical warning rules include: For general hidden dangers, early warning information will be sent to power users via SMS push, and power users will be reminded to make corrections within the specified time; In the event of a major hidden danger, early warning information will be sent to power users and power management departments via SMS push, APP push, and phone notifications, requiring power users to immediately stop the operation of related equipment, make rectifications within a specified time, and take preventive measures; The corresponding graded lighting warning rules include: Green / blue light is always on: 90 points < lighting score ≤ 100 points, no general hidden dangers or major hidden dangers, and no electricity use behavior that violates the contract; Yellow light is constantly flashing: 60 points < lighting score < 90 points, hidden dangers and defects are being rectified, there is a breach of contract in electricity use, and the process has been initiated; Red light is always on: 0 points < lighting score ≤ 60 points, hidden dangers and defects have not been rectified, there is a breach of contract in electricity use, and the process has not been initiated; All lights are off: monitoring interruption or power outage; The rectification management module generates rectification plans based on the safety hazard identification results, including: determining the type of safety hazard and extracting the corresponding rectification plan from the database according to the safety hazard type.

[0014] The above also includes a visual interaction unit for information interaction with power users.

[0015] The present invention collects the electrical equipment operating parameters, environmental parameters and user electricity usage behavior data of important power users, introduces deep learning to establish a hidden danger analysis model, analyzes the electrical equipment operating parameters, environmental parameters and user electricity usage behavior data of important power users, accurately identifies the type and severity of safety hazards, effectively solves the problem of inadequate hidden danger detection of important power users, significantly improves the accuracy, timeliness and comprehensiveness of hidden danger detection, and provides strong guarantee for the safe and stable operation of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Attachment Figure 1 This is a flow chart of the hidden danger detection and dynamic control method provided by the present invention.

[0017] Attachment Figure 2 This is a flow chart of the method for constructing a hidden danger analysis model provided by the present invention.

[0018] Attachment Figure 3 Schematic diagram of the network structure of the initial machine learning model provided by the present invention.

[0019] Attachment Figure 4 This is a flow chart of the method for obtaining real-time basic data to be analyzed provided by the present invention.

[0020] Attachment Figure 5 Schematic diagram of the network structure of edge computing provided by the present invention.

[0021] Attachment Figure 6 This is a schematic diagram of the structure of a hidden danger detection and dynamic management and control system provided by the present invention.

[0022] Attachment Figure 7 This is a schematic diagram of the hidden danger identification unit structure provided by the present invention.

[0023] Attachment Figure 8 This is a structural diagram of the data acquisition unit provided by the present invention.

[0024] Attachment Figure 9 This is a schematic diagram of the structure of the dynamic control unit provided by the present invention.

[0025] Attachment Figure 10 This is a schematic diagram of the structure of another hidden danger detection and dynamic control system provided by the present invention. DETAILED DESCRIPTION

[0026] The present invention is not limited to the following embodiments, and specific implementation methods can be determined based on the technical solutions of the present invention and actual conditions.

[0027] The present application is further described below with reference to the embodiments and accompanying drawings: Example 1: As shown in the attached Figure 1 As shown, the embodiment of the present invention discloses a method for troubleshooting and dynamically controlling hidden dangers of important power users, including: Step S110, obtaining real-time basic data to be analyzed, wherein the real-time basic data to be analyzed includes electrical equipment operating parameters, environmental parameters, and user electricity usage behavior data; The above-mentioned electrical equipment operating parameters include voltage, current, power factor, and temperature; the environmental parameters include humidity, temperature, and smoke concentration; and the user electricity usage behavior data includes electricity load curves and equipment start and stop times.

[0028] Step S120: Input the real-time basic data to be analyzed into the hidden danger analysis model to obtain safety hidden danger identification results. The hidden danger analysis model is obtained by training an initial machine learning model using a number of samples. The initial machine learning model includes a long short-term memory network and a convolutional neural network. Each sample includes electrical equipment operating parameters, environmental parameters, user electricity usage behavior data, and corresponding safety hidden danger type and severity identification information. The above-mentioned safety hazard types may include equipment temperature slightly higher than the normal range, a slight decrease in power factor, transformer oil temperature too high close to the tripping threshold, serious line overload, fire smoke alarm, line short circuit causing sparks, etc. General hazards include equipment temperature slightly higher than the normal range, a slight decrease in power factor, etc. Major hazards include transformer oil temperature too high close to the tripping threshold, serious line overload, fire smoke alarm, line short circuit causing sparks, etc.

[0029] Step S130: Perform graded warning based on the safety hazard identification results.

[0030] When performing graded warnings based on the safety hazard identification results, the corresponding graded warning rules include: For general hidden dangers, early warning information will be sent to power users via SMS push notifications, reminding them to make corrections within a specified timeframe. The information will include a detailed description of the hidden danger, the potential impact, and recommended corrective measures, with users reminded to make corrections within seven working days. For major hidden dangers, early warning information will be sent to power users and power management departments through SMS push, APP push, and telephone notifications, requiring power users to immediately stop the operation of related equipment, make rectifications within the specified time, and take preventive measures; the content of the SMS text message concisely and clearly explains the key information and severity of the hidden danger, requiring users to immediately stop the operation of related equipment and complete rectifications within 3 working days. Further professional personnel can be arranged for on-site guidance.

[0031] The present invention discloses a method for detecting and dynamically controlling hidden dangers of important power users. The method collects the operating parameters, environmental parameters and electricity usage behavior data of electrical equipment of important power users, introduces deep learning to establish a hidden danger analysis model, analyzes the operating parameters, environmental parameters and electricity usage behavior data of electrical equipment of important power users, accurately identifies the type and severity of safety hazards, effectively solves the problem of inadequate detection of hidden dangers of important power users, significantly improves the accuracy, timeliness and comprehensiveness of hidden danger detection, realizes the "early detection and early treatment" of hidden dangers of important power users, and provides a strong guarantee for the safe and stable operation of the power system.

[0032] Example 2: As shown in the attached Figure 2 As shown, the embodiment of the present invention is a further optimization of the above embodiment, wherein the process of constructing the hidden danger analysis model includes: Step S210: Acquire a number of historical hidden danger data, identify the type and severity of the safety hidden danger corresponding to each historical hidden danger data, obtain a sample set, and divide the sample set into a training sample set and a test sample set, wherein the historical hidden danger data includes electrical equipment operating parameters, environmental parameters, and user electricity usage behavior data; The above historical hidden danger data is not limited to hidden danger data when the equipment fails, but also includes data when the equipment is normal; and the ratio used when dividing the sample set can be set according to actual needs. In this embodiment, it can be set to 7:3.

[0033] Step S220: Train the initial machine learning model using the training sample set. End the training when the training stop condition is met to obtain a hidden danger analysis model. The initial machine learning model includes a long short-term memory network, a convolutional neural network, a fully connected layer, and a classifier. The long short-term memory network extracts long-term dependency features of the input data, and the convolutional neural network extracts spatial features of the input data. The long-term dependency features and spatial features are fused, and the fused features are classified using the fully connected layer and the classifier to obtain the type and severity of the safety hazard. The network structure of the above initial machine learning model is shown in the attached Figure 3 As shown in the figure; the fusion of long-term dependency features and spatial features can be, but is not limited to, convolution and splicing; the long short-term memory network extracts the long-term dependency features of the input data, that is, extracts the long-term dependency relationship in the time series of the input data, such as analyzing the changing trends of voltage and current over a period of time, and judging whether there is excessive voltage fluctuation or abnormal increase in current; the convolutional neural network extracts the spatial features of the input data, such as extracting the correlation features between the parameters of different electrical equipment, and judging whether there are safety hazards caused by mutual influence between the equipment; the fused features are classified using the fully connected layer and the classifier, and the types of safety hazards obtained may include equipment overheating, line short circuit, low power factor, etc.

[0034] Step S230 , using the test set to test the trained hidden danger analysis model, optimize the model parameters of the hidden danger analysis model, and output the hidden danger analysis model that meets the test evaluation requirements.

[0035] The back propagation algorithm can be used to adjust the model parameters of the hidden danger analysis model mentioned above.

[0036] Example 3: As shown in the attached Figure 4 As shown, the embodiment of the present invention is a further optimization of the above embodiment, wherein obtaining real-time basic data to be analyzed includes: Step S310: Acquire real-time basic data from important power users, where the real-time basic data includes electrical equipment operating parameters, environmental parameters, and user electricity usage behavior data; The above-mentioned electrical equipment operating parameters include voltage, current, power factor, and temperature; the environmental parameters include humidity, temperature, and smoke concentration; and the user electricity usage behavior data includes electricity load curves and equipment start and stop times.

[0037] In step S320, edge computing is introduced to screen and process the real-time basic data to establish the real-time basic data to be analyzed. The edge computing includes feature extraction through different convolution kernel combinations, data dimensionality reduction using maximum pooling, and integration of the convolution results and pooling results using a fully connected layer.

[0038] The network structure of the above edge computing is shown in the attached Figure 5 As shown in the figure, different convolution kernel combinations are used for feature extraction. The specific convolution layer uses a 3×3 convolution kernel to extract features, such as extracting the fluctuation characteristics of voltage and current data, and the changing trend characteristics of temperature data. Maximum pooling is used to reduce the dimension of data, reduce the amount of data, and retain important feature information to improve data processing efficiency. The fully connected layer is used to integrate the convolution results and pooling results, filter out invalid data, and select key data.

[0039] Example 4: This embodiment of the present invention is a further optimization of the above embodiment, which also includes generating a rectification plan based on the safety hazard identification results, including: Identify the type of safety hazard; Extract the corresponding rectification plan from the database according to the type of safety hazard.

[0040] The above database can store in advance the corresponding rectification plans for each type of safety hazard, and update them regularly. For example, for the hazard of equipment overheating, the rectification plan may include cleaning the equipment's heat dissipation channels, checking whether the cooling system is working properly, replacing aging cooling fans, and other measures; for the hazard of line short circuit, the rectification plan includes checking whether the line insulation layer is damaged, finding the short circuit point and repairing it, and performing insulation testing on the line.

[0041] It should also be noted that in order to track the progress of rectification in real time, electricity users are required to upload photos, videos and other materials of the rectification process through the State Grid APP online, and record the rectification time, rectification measures and rectification results.

[0042] Example 5: As shown in the attached Figure 6 As shown, the embodiment of the present invention discloses a system for detecting and dynamically controlling hidden dangers of important power users, including: A data acquisition unit acquires real-time basic data to be analyzed, wherein the real-time basic data to be analyzed includes electrical equipment operating parameters, environmental parameters, and user electricity consumption behavior data; The hidden danger identification unit inputs the real-time basic data to be analyzed into the hidden danger analysis model to obtain the safety hidden danger identification results. The hidden danger analysis model is obtained by training an initial machine learning model using several samples. The initial machine learning model includes a long short-term memory network and a convolutional neural network. Each sample includes electrical equipment operating parameters, environmental parameters, user electricity usage behavior data, and corresponding safety hidden danger type and severity identification information; Dynamic control unit, which provides graded warnings based on safety hazard identification results; The database stores various types of system data. Specifically, it can store collected data, hidden danger analysis model training data, hidden danger inspection history records, and rectification records.

[0043] The aforementioned hidden danger identification unit, dynamic control unit, and database can all be simultaneously installed in electronic devices such as computers and servers, or independently installed in different devices. In this embodiment, the data acquisition unit, hidden danger identification unit, and dynamic control unit form a closed-loop system for troubleshooting and dynamically controlling hidden dangers for important power users. This system is capable of performing real-time and accurate hidden danger inspections for important power users, eliminating the problems of low efficiency, significant subjective influence, and lagging management associated with the original inspection method that relied on manual inspections and periodic testing.

[0044] Example 6: As shown in the attached Figure 7 As shown, the embodiment of the present invention is a further optimization of the above embodiment, wherein the hidden danger identification unit includes: Model building subunit, including: The sample acquisition module acquires a number of historical hidden danger data, identifies the type and severity of the safety hazard corresponding to each historical hidden danger data, obtains a sample set, and divides the sample set into a training sample set and a test sample set. The historical hidden danger data includes electrical equipment operating parameters, environmental parameters, and user electricity usage behavior data; The model training module uses the training sample set to train the initial machine learning model. When the training stop conditions are met, the training ends and a hidden danger analysis model is obtained. The initial machine learning model includes a long short-term memory network, a convolutional neural network, a fully connected layer, and a classifier. The long short-term memory network extracts the long-term dependency features of the input data, and the convolutional neural network extracts the spatial features of the input data. The long-term dependency features and spatial features are fused, and the fused features are classified using a fully connected layer and a classifier to obtain the type and severity of the safety hazard. The model testing module uses the test set to test the trained hidden danger analysis model, optimizes the model parameters of the hidden danger analysis model, and outputs a hidden danger analysis model that meets the test and evaluation requirements; The hidden danger identification sub-unit inputs the real-time basic data to be analyzed into the hidden danger analysis model to obtain the safety hidden danger identification results.

[0045] The functional steps of the above modules are the same as those in the above embodiment and will not be repeated here.

[0046] Example 7: As shown in the attached Figure 8 As shown, the embodiment of the present invention is a further optimization of the above embodiment, wherein the data acquisition unit includes: The acquisition module, including smart meters, load monitoring devices and various sensors, obtains real-time basic data from important power users, including electrical equipment operating parameters, environmental parameters and user electricity usage behavior data.

[0047] These types of sensors include: The electrical parameter sensor group includes a voltage sensor, a current sensor, a power factor sensor, and a temperature sensor, which are used to collect the voltage, current, power factor, and temperature parameters of electrical equipment respectively.

[0048] The voltage sensor uses an electromagnetic induction voltage transformer, and the current sensor uses a high-precision Hall effect current sensor. These sensors are installed on key electrical equipment such as power distribution cabinets and transformers. They can collect voltage and current data in real time, with a collection frequency set to once per second, ensuring that dynamic changes in electrical equipment operation are captured. The power factor sensor monitors the power factor in real time by calculating the phase difference between voltage and current. The temperature sensor uses a fiber optic temperature sensor, which enables non-contact temperature measurement of transformer windings, cable connectors, and other parts, avoiding measurement errors and safety hazards caused by contact.

[0049] The environmental parameter sensor group includes a humidity sensor, a temperature sensor, and a smoke concentration sensor, which are used to collect the humidity, temperature, and smoke concentration parameters of the operating environment of the power equipment respectively.

[0050] The humidity sensor uses a capacitive humidity sensor, the temperature sensor uses a digital temperature sensor, and the smoke concentration sensor uses an infrared beam smoke detector. These sensors are distributed throughout the operating environment of power equipment, such as distribution rooms and cable trenches, to monitor changes in humidity, temperature, and smoke concentration in real time. If environmental parameters exceed normal ranges, they may affect the safe operation of power equipment.

[0051] The above-mentioned smart meters and load monitoring devices are used to collect user electricity usage behavior data and generate electricity load curves and equipment start and stop time data.

[0052] Smart meters and load monitoring devices are installed at the power user's incoming line end. Smart meters can accurately measure the user's electricity consumption and collect electricity consumption time information. Load monitoring devices generate electricity load curves by analyzing current data and record the start and stop time of the equipment. Through these data, the user's electricity consumption behavior patterns can be analyzed to detect abnormal electricity consumption, such as potential hidden dangers such as long-term overload operation of equipment.

[0053] The edge computing module extracts features through different combinations of convolution kernels, uses maximum pooling for data dimensionality reduction, and uses a fully connected layer to integrate the convolution results and pooling results to obtain real-time basic data to be analyzed.

[0054] Example 8: As shown in the attached Figure 9 As shown, the embodiment of the present invention is a further optimization of the above embodiment, wherein the dynamic control unit includes: A graded warning module provides graded warnings and light warnings based on the results of potential safety hazards identification; The corresponding hierarchical warning rules include: For general hidden dangers, early warning information will be sent to power users via SMS push, and power users will be reminded to make corrections within the specified time; In the event of a major hidden danger, early warning information will be sent to power users and power management departments via SMS push, APP push, and phone notifications, requiring power users to immediately stop the operation of related equipment, make rectifications within a specified time, and take preventive measures; The corresponding graded lighting warning rules include: Green / blue light is always on: 90 points < lighting score ≤ 100 points, no general hidden dangers or major hidden dangers, and no electricity use behavior that violates the contract; Yellow light is constantly flashing: 60 points < lighting score < 90 points, hidden dangers and defects are being rectified, there is a breach of contract in electricity use, and the process has been initiated; Red light is always on: 0 points < lighting score ≤ 60 points, hidden dangers and defects have not been rectified, there is a breach of contract in electricity use, and the process has not been initiated; All lights are off: monitoring interruption or power outage; The rectification management module generates rectification plans based on the safety hazard identification results, including: determining the type of safety hazard and extracting the corresponding rectification plan from the database according to the safety hazard type.

[0055] Example 9: As shown in the attached Figure 10 As shown, the embodiment of the present invention discloses a system for detecting and dynamically controlling hidden dangers of important power users, including: Visual interaction unit, which exchanges information with power users; The visual interaction unit interacts with power users through channels such as the State Grid app and online customer service. Power users can use the app to query potential hazard warning information, rectification plans, and rectification progress. They can also provide feedback to the interaction unit regarding any issues encountered during the rectification process. The interaction unit's online customer service staff promptly answer user questions and provide technical support, such as guidance on how to properly use power equipment and perform basic equipment maintenance. The interaction unit also collects user feedback and suggestions to inform system optimization and improvement.

[0056] A data acquisition unit acquires real-time basic data to be analyzed, wherein the real-time basic data to be analyzed includes electrical equipment operating parameters, environmental parameters, and user electricity consumption behavior data; The hidden danger identification unit inputs the real-time basic data to be analyzed into the hidden danger analysis model to obtain the safety hidden danger identification results. The hidden danger analysis model is obtained by training an initial machine learning model using several samples. The initial machine learning model includes a long short-term memory network and a convolutional neural network. Each sample includes electrical equipment operating parameters, environmental parameters, user electricity usage behavior data, and corresponding safety hidden danger type and severity identification information; The dynamic management and control unit provides graded warnings based on the results of safety hazard identification.

[0057] The above content is only a specific implementation method of the present application, which has strong adaptability and implementation effect, but the protection scope of the present application is not limited to this. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed in this application, which should be included in the protection scope of the present application. Therefore, equivalent changes made according to the claims of this application are still within the scope covered by this application.

Claims

1. A method for troubleshooting and dynamic management of hidden dangers of important power users, characterized in that: include: Obtaining real-time basic data to be analyzed, where the real-time basic data to be analyzed includes electrical equipment operating parameters, environmental parameters, and user electricity usage behavior data; Input the real-time basic data to be analyzed into the hidden danger analysis model to obtain the safety hidden danger identification results. The hidden danger analysis model is obtained by training an initial machine learning model using several samples. The initial machine learning model includes a long short-term memory network and a convolutional neural network. Each sample includes electrical equipment operating parameters, environmental parameters, user electricity usage behavior data, and corresponding safety hidden danger type and severity identification information; Provide graded warnings based on safety hazard identification results.

2. The method for troubleshooting and dynamically controlling hidden dangers of important power users according to claim 1 is characterized in that: The process of constructing the hidden danger analysis model includes: Obtaining a number of historical hidden danger data, identifying the type and severity of the safety hidden danger corresponding to each historical hidden danger data, obtaining a sample set, and dividing the sample set into a training sample set and a test sample set, wherein the historical hidden danger data includes electrical equipment operating parameters, environmental parameters, and user electricity usage behavior data; The initial machine learning model is trained using the training sample set. Training ends when the training stopping conditions are met to obtain a hidden danger analysis model. The initial machine learning model includes a long short-term memory network, a convolutional neural network, a fully connected layer, and a classifier. The long short-term memory network extracts the long-term dependency features of the input data, and the convolutional neural network extracts the spatial features of the input data. The long-term dependency features and spatial features are fused, and the fused features are classified using a fully connected layer and a classifier to obtain the type and severity of the safety hazard. The trained hidden danger analysis model is tested using the test set to optimize the model parameters of the hidden danger analysis model and output a hidden danger analysis model that meets the test and evaluation requirements.

3. The method for troubleshooting and dynamically controlling hidden dangers of important power users according to claim 1 or 2, characterized in that: The obtaining of real-time basic data to be analyzed includes: Obtain real-time basic data from important power users, including electrical equipment operating parameters, environmental parameters, and user electricity usage behavior data; Edge computing is introduced to screen and process real-time basic data to establish real-time basic data to be analyzed. Edge computing includes feature extraction through different convolution kernel combinations, data dimensionality reduction using maximum pooling, and integration of convolution results and pooling results using a fully connected layer.

4. The method for troubleshooting and dynamically controlling hidden dangers of important power users according to any one of claims 1 to 3, characterized in that: When performing graded warning based on the safety hazard identification results, the corresponding graded warning rules include: For general hidden dangers, early warning information will be sent to power users via SMS push, and power users will be reminded to make corrections within the specified time; For major hidden dangers, early warning information will be sent to power users and power management departments through SMS push, APP push, and telephone notifications, and power users will be required to immediately stop the operation of related equipment, make rectifications within the specified time, and take preventive measures.

5. The method for troubleshooting and dynamically controlling hidden dangers of important power users according to any one of claims 1 to 3, characterized in that: It also includes generating rectification plans based on the safety hazard identification results, including: Identify the type of safety hazard; Extract the corresponding rectification plan from the database according to the type of safety hazard.

6. A system for detecting and dynamically controlling hidden dangers of important power users using the method according to any one of claims 1 to 5, characterized in that: include: A data acquisition unit acquires real-time basic data to be analyzed, wherein the real-time basic data to be analyzed includes electrical equipment operating parameters, environmental parameters, and user electricity consumption behavior data; The hidden danger identification unit inputs the real-time basic data to be analyzed into the hidden danger analysis model to obtain the safety hidden danger identification results. The hidden danger analysis model is obtained by training an initial machine learning model using several samples. The initial machine learning model includes a long short-term memory network and a convolutional neural network. Each sample includes electrical equipment operating parameters, environmental parameters, user electricity usage behavior data, and corresponding safety hidden danger type and severity identification information; The dynamic management and control unit provides graded warnings based on the results of safety hazard identification.

7. The system for detecting and dynamically controlling hidden dangers of important power users according to claim 6 is characterized in that: The hidden danger identification unit includes: Model building subunit, including: The sample acquisition module acquires a number of historical hidden danger data, identifies the type and severity of the safety hazard corresponding to each historical hidden danger data, obtains a sample set, and divides the sample set into a training sample set and a test sample set. The historical hidden danger data includes electrical equipment operating parameters, environmental parameters, and user electricity usage behavior data; The model training module uses the training sample set to train the initial machine learning model. When the training stop conditions are met, the training ends and a hidden danger analysis model is obtained. The initial machine learning model includes a long short-term memory network, a convolutional neural network, a fully connected layer, and a classifier. The long short-term memory network extracts the long-term dependency features of the input data, and the convolutional neural network extracts the spatial features of the input data. The long-term dependency features and spatial features are fused, and the fused features are classified using a fully connected layer and a classifier to obtain the type and severity of the safety hazard. The model testing module uses the test set to test the trained hidden danger analysis model, optimizes the model parameters of the hidden danger analysis model, and outputs a hidden danger analysis model that meets the test and evaluation requirements; The hidden danger identification sub-unit inputs the real-time basic data to be analyzed into the hidden danger analysis model to obtain the safety hidden danger identification results.

8. The system for detecting and dynamically controlling hidden dangers of important power users according to claim 6 or 7 is characterized in that: The data acquisition unit includes: The acquisition module, including smart meters, load monitoring devices, and various sensors, acquires real-time basic data from important power users, including electrical equipment operating parameters, environmental parameters, and user electricity usage behavior data; The edge computing module extracts features through different combinations of convolution kernels, uses maximum pooling for data dimensionality reduction, and uses a fully connected layer to integrate the convolution results and pooling results to obtain real-time basic data to be analyzed.

9. The system for detecting and dynamically controlling hidden dangers of important power users according to any one of claims 6 to 8, characterized in that: The dynamic control unit includes: A graded warning module provides graded warnings and light warnings based on the results of potential safety hazards identification; The corresponding hierarchical warning rules include: For general hidden dangers, early warning information will be sent to power users via SMS push, and power users will be reminded to make corrections within the specified time; In the event of a major hidden danger, early warning information will be sent to power users and power management departments via SMS push, APP push, and phone notifications, requiring power users to immediately stop the operation of related equipment, make rectifications within a specified time, and take preventive measures; The corresponding graded lighting warning rules include: Green / blue light is always on: 90 points < lighting score ≤ 100 points, no general hidden dangers or major hidden dangers, and no electricity use behavior that violates the contract; Yellow light is constantly flashing: 60 points < lighting score < 90 points, hidden dangers and defects are being rectified, there is a breach of contract in electricity use, and the process has been initiated; Red light is always on: 0 points < lighting score ≤ 60 points, hidden dangers and defects have not been rectified, there is a breach of contract in electricity use, and the process has not been initiated; All lights are off: monitoring interruption or power outage; The rectification management module generates rectification plans based on the safety hazard identification results, including: determining the type of safety hazard and extracting the corresponding rectification plan from the database according to the safety hazard type.

10. The system for detecting and dynamically controlling hidden dangers of important power users according to any one of claims 6 to 9, characterized in that: It also includes a visual interaction unit for information exchange with electricity users.

Images