Electric power safety tool detection system and detection method thereof

Abstract

The invention discloses an electric power safety tool detection system and a detection method thereof, and relates to the technical field of electric power safety, and the system comprises the following components: a non-contact sensor array which integrates various non-contact sensors such as an infrared sensor, a laser range finder and an ultrasonic flaw detector, the infrared, laser and ultrasonic non-contact sensor arrays are respectively used for detecting the insulating property, the mechanical strength and the wear degree of the tool, and the multi-dimensional and high-precision automatic detection of the insulating property, the mechanical strength and the wear degree of the electric power safety tool is realized by adopting the infrared, laser and ultrasonic non-contact sensor arrays; according to the method, secondary damage possibly caused by traditional contact type detection to the tool is effectively avoided, particularly when the tool in the high-pressure and high-temperature extreme environment is treated, the safety of the detection process is remarkably improved through the non-contact characteristic of the method, subtle changes of the tool can be captured more accurately through non-contact type detection, and the detection accuracy of the tool is improved. Therefore, the detection accuracy and reliability are obviously improved.

Description

Technical Field

[0001] The present invention relates to the technical field of electric power safety, and in particular to an electric power safety tool detection system and a detection method thereof. Background Art

[0002] With the rapid development of the power industry, power safety tools, as important tools to ensure the stable operation of the power system and personnel safety, their performance status is directly related to the safety and reliability of the power grid operation. Power safety tools will experience various harsh environments during use. These environmental factors can easily lead to problems such as decreased insulation performance, weakened mechanical strength, and increased wear of the tools. Therefore, it is particularly important to conduct regular and efficient inspection and maintenance of power safety tools.

[0003] Although traditional detection technologies such as visual inspection and manual measurement can evaluate the status of tools to a certain extent, these methods have many limitations. First, most traditional methods rely on manual operation, which is not only inefficient but also easily affected by human factors, resulting in inconsistent detection results. Second, for tools under high pressure and high temperature extreme environments, traditional contact detection methods such as direct measurement and tapping tests are not only difficult to implement, but may also cause secondary damage to the tools, increasing safety risks. More importantly, traditional detection technology is often limited to single-dimensional detection, focusing on insulation resistance values ​​and external wear, and ignoring the monitoring of subtle changes inside the tools. Subtle changes are often a precursor to safety accidents.

[0004] Therefore, in response to the existing problems, an electric power safety tool detection system and its detection method are developed, which not only significantly improves the accuracy and reliability of detection, but also greatly enhances the detection and diagnosis capabilities of electric power safety tools, providing more solid technical support for the safe and stable operation of the power system. Summary of the Invention

[0005] The purpose of the present invention is to make up for the shortcomings of the existing technology and provide an electric power safety tool detection system and detection method. It can realize multi-dimensional and high-precision automatic detection of tools by integrating infrared, laser, and ultrasonic non-contact sensor arrays. It not only avoids the secondary damage and safety risks that may be caused by traditional contact detection, but also greatly improves the accuracy and comprehensiveness of detection. Combined with advanced machine learning algorithms, it can perform in-depth analysis of complex multi-source data, realize comprehensive evaluation of tool performance and precise fault location.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: On the one hand, a power safety tool detection system includes the following components:

[0007] Non-contact sensor array: Integrates infrared sensors, laser rangefinders, ultrasonic flaw detectors and other non-contact sensors to detect the insulation performance, mechanical strength and wear degree of tools;

[0008] Data acquisition and processing module: responsible for receiving data collected by various sensors, performing preprocessing, denoising, filtering, and extracting key feature information;

[0009] Multi-dimensional analysis module: Based on machine learning and deep learning algorithms, it performs multi-dimensional analysis on the processed data to accurately determine the insulation status, strength level and wear of the tools and generate a test report;

[0010] Human-computer interaction interface: Provides an intuitive operation interface to display test results, historical records and warning information, supports user query and report export functions, and enhances user experience.

[0011] Furthermore, the non-contact sensor array detects insulation performance and leakage current analysis through infrared thermal imaging, sets infrared thermal imager parameters, performs infrared thermal imaging scanning on the tool, obtains the infrared radiation intensity of each point, converts the received infrared radiation intensity data into a temperature value, forms a temperature distribution image of the tool surface, and calculates the average temperature of the entire image. The formula is: Among them, T ij is the temperature of the point with coordinates (i, j) in the image, N and M are the number of rows and columns of the image respectively. The standard deviation σ of the temperature is calculated as follows: A threshold of 3σ is set. If the difference from the average temperature exceeds the threshold, it is considered to be a temperature anomaly area. The temperature gradient of the temperature anomaly area is calculated, and the possibility and location of leakage current are evaluated by combining Joule's law and heat conduction equation.

[0012] Furthermore, the non-contact sensor array detects mechanical strength using a laser rangefinder, continuously measuring the tool at a preset measurement frequency. The laser rangefinder emits a laser beam to illuminate the surface of the tool, and calculates the distance L by measuring the time difference Δt between the emitted light and the reflected light. The distance calculation formula is: Where c is the speed of light, calculate the distance change between adjacent measurement points: ΔL i =L i+1 -L i , calculate the average value of the distance change: Calculate the standard deviation of the distance variation: By analyzing the time series curve of the distance change, the deformation trend of the tool during force and vibration is observed. Combined with the mechanical properties and geometric shape of the materials used in the tool, the reliability of the mechanical strength of the tool under the current force and vibration conditions is evaluated.

[0013] Furthermore, the non-contact sensor array detects wear and cracks using an ultrasonic flaw detector. An ultrasonic flaw detector probe that matches the material of the tool is selected, and multiple detection areas are selected on the tool surface. The ultrasonic flaw detector emits a series of short ultrasonic pulses into the tool. The pulse propagation velocity v is determined by the material properties and is expressed as: Where E is the elastic modulus of the material, ρ is the material density, and the reflected ultrasonic signal is received. The received reflected signal is analyzed and its waveform characteristics are extracted. The defect position L is calculated by the time-of-flight method. The formula is: Where t is the time difference between the transmitted pulse and the received reflected pulse. The amplitude change ΔA and phase change Δφ of the reflected signal, combined with the beam diffusion characteristics of the probe, can be used to identify the location, size, and shape of internal cracks and surface wear. The degree of wear can be assessed by combining the acoustic characteristics of the tool material and the tool's usage history.

[0014] Furthermore, the data acquisition and processing module monitors and receives data streams from multiple non-contact sensors in real time through a serial communication interface, converts the received raw data from a sensor-specific format into a unified data format, filters the data, applies a feature extraction algorithm to process the filtered data, extracts temperature anomaly areas, deformation degrees, and crack sizes, and stores the feature-extracted data in a database.

[0015] Furthermore, the multi-dimensional analysis module receives relevant data from the data acquisition and processing module, classifies and organizes the data according to different tool types, inspection items and time series to form a structured data set, selects support vector machines to judge the insulation status, adopts random forests to evaluate the strength level, uses convolutional neural networks to analyze the wear conditions, integrates the output results of multiple models, conducts multi-dimensional analysis of tools, generates detailed inspection reports based on the results of analysis and judgment, and uses charts to display the data.

[0016] Furthermore, the multi-dimensional analysis module selects the support vector machine to judge the insulation status, collects the data related to the insulation performance of the tool, and expresses it as a feature vector x i , the corresponding label y i ∈{-1,1}, -1 means good insulation, 1 means there is a problem, and the Gaussian kernel function is selected as the kernel function of the model, and its formula is: K(x i ,x j )=exp(-γ||x i -x j || 2 ), where γ is the parameter of the kernel function. The goal of SVM is to maximize the distance between the two types of data points and the hyperplane. The optimization problem is expressed as Where C is the penalty parameter, ξ i is the slack variable, φ(x i ) is a function that maps input data to a high-dimensional feature space, w is the normal vector of the hyperplane, b is the intercept of the hyperplane, and the Lagrange multiplier method is used to transform the above optimization problem into its dual problem and solve the obtained α i , calculate w and b, for any support vector x j The final decision function is: f(x) = sign For new tools insulation performance data x new Substitute it into the decision function to get the predicted label, thereby judging whether its insulation status is good or has problems.

[0017] Furthermore, the multidimensional analysis module uses random forest to evaluate the intensity level, collects relevant data of tools and represents it as a feature vector X, and the corresponding intensity level label is Y. Random sampling with replacement is performed from the original data set to create multiple bootstrap sample sets. For each bootstrap sample set, a decision tree is constructed, m features are randomly selected from the feature set, and the optimal splitting feature and splitting point are selected according to the Gini index criterion. The Gini index formula is: Where K is the number of categories, p k,t is the proportion of samples belonging to category k in node t. Repeatedly construct multiple decision trees to form a random forest. For new tool data, input it into each decision tree in the random forest to predict X new , obtain prediction results of multiple intensity levels, and integrate multiple prediction results by voting to obtain the final intensity level evaluation.

[0018] Furthermore, the multi-dimensional analysis module uses a convolutional neural network to analyze the wear condition, collects tool wear image data, marks the area and degree of wear, receives the wear image data, and uses a convolution kernel to perform a convolution operation on the input image to extract features. The convolution operation is expressed as: Among them, Z ij is the element of the output feature map, W mn is the element of the convolution kernel, X i+m,j+n is the element of the input image, b is the bias term, maximum pooling is used to reduce the dimension of the feature map, the extracted features are mapped to the final output category through the fully connected layer, and the cross entropy loss function is used as the model loss function, which is expressed as: Where N is the number of samples, y i is the true label, It is the prediction probability. The stochastic gradient descent algorithm is used to update the weights and biases of the network. The new tool image to be analyzed is input into the trained convolutional neural network to obtain the prediction result of the wear degree.

[0019] On the other hand, a method for detecting electric safety tools is provided, wherein the specific steps of the method are as follows:

[0020] Data collection and processing: The power safety tools to be tested are placed in the testing area, and the multi-sensor fusion detection system is activated. Each sensor simultaneously collects multi-dimensional data from the tools. Advanced image processing and signal analysis algorithms are used to pre-process, extract features, and perform pattern recognition on the collected data.

[0021] Multi-dimensional inspection and evaluation: Based on the processed data, comprehensive, multi-dimensional inspection and evaluation of tools are carried out, including:

[0022] Insulation performance testing: Analyze the temperature distribution on the surface of the tool through infrared images, and combine laser scanning data to determine whether the insulation layer is damaged, aged, or falling off;

[0023] Mechanical strength assessment: Use laser sensors to accurately measure the size and shape changes of tools, and use ultrasonic waves to detect internal cracks and defects to comprehensively assess their mechanical strength;

[0024] Wear degree detection: Use ultrasonic sensors to detect wear marks on the surface and internal wear of tools to evaluate their service life;

[0025] Automatic classification and labeling: According to the test results, the tools are automatically classified into qualified, repairable and scrapped levels, and their test status and results are recorded;

[0026] Data storage and report generation: The test data, results and classification information are stored in the database, allowing users to query historical records at any time. The system generates detailed test reports based on needs and supports exporting in multiple formats.

[0027] Compared with the existing technology, this power safety tool detection system and detection method have the following advantages:

[0028] Beneficial effects:

[0029] 1. The infrared, laser, and ultrasonic non-contact sensor arrays used in this invention enable multi-dimensional, high-precision automatic detection of the insulation performance, mechanical strength, and wear level of power safety tools. This method effectively avoids the secondary damage that traditional contact detection may cause to tools. In particular, when dealing with tools in extreme high-voltage and high-temperature environments, its non-contact characteristics significantly improve the safety of the detection process. Through non-contact detection, subtle changes in tools can be more accurately captured, thereby significantly improving the accuracy and reliability of detection.

[0030] 2. The multi-dimensional analysis module of the present invention integrates a variety of machine learning algorithms, which can perform in-depth analysis and comprehensive judgment on the data collected by different sensors. This multi-dimensional data processing method not only realizes a comprehensive evaluation of the performance of tools, but also can improve the accuracy and comprehensiveness of diagnosis through mutual verification and supplementation between models. This combination of multi-dimensional detection and analysis significantly enhances the detection and diagnosis capabilities of power safety tools, providing a strong guarantee for the safe operation of the power system.

[0031] Other advantages, objects and features of the present invention will be described in part in the following description and, in part, will be apparent to those skilled in the art based on an examination of the following or may be learned from the practice of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort.

[0033] Figure 1 This is a process operation diagram of an electric safety tool detection system;

[0034] Figure 2 The present invention is a flow chart of a method for detecting electric safety tools. DETAILED DESCRIPTION

[0035] The following is a clear and complete description of the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0036] Example 1:

[0037] This embodiment of the invention relates to a detection system for electric safety tools. This patent integrates advanced technologies such as infrared thermal imagers, laser scanners, and ultrasonic sensors to achieve comprehensive and accurate detection of the insulation performance, mechanical strength, and degree of wear of tools. By synchronously collecting multi-dimensional data and using advanced image processing and signal analysis algorithms for preprocessing, feature extraction, and pattern recognition, it can automatically analyze the status of tools and accurately classify them into qualified, repairable, and scrapped levels.

[0038] The non-contact sensor array uses infrared sensors to detect the temperature distribution on the surface of tools, identify abnormal temperature areas, and indirectly evaluate the insulation performance. It uses a laser rangefinder to measure tiny deformations and distance changes on the surface of tools to evaluate their mechanical strength and degree of deformation. The ultrasonic flaw detector uses the propagation characteristics of ultrasonic waves in materials to detect cracks and defects inside tools and evaluate their degree of wear and structural integrity.

[0039] The data acquisition and processing module monitors the data streams from infrared sensors, laser rangefinders, and ultrasonic flaw detectors in real time through a serial communication interface. It converts the received raw data from the sensor-specific format into a unified data format, removes high-frequency noise and random interference in the data, uses digital filtering technology to smooth the data and reduce noise interference, and applies a feature extraction algorithm to process the filtered data to extract key features such as temperature anomaly areas, deformation degree, and crack size. The preprocessed, denoised, filtered, and feature-extracted data is stored in a database.

[0040] The multi-dimensional analysis module receives relevant data from the data acquisition and processing module, classifies and organizes the data according to the tool type, inspection items and time series to form a structured data set, selects the support vector machine to judge the insulation status, analyzes the extracted temperature characteristics through the trained model, evaluates the insulation performance of the tool, adopts random forest to evaluate the strength level, comprehensively evaluates the mechanical strength of the tool based on the degree of deformation and laser ranging data, uses convolutional neural network to analyze the wear condition, uses ultrasonic flaw detection data, identifies internal cracks and wear degree through deep learning model, integrates the output results of multiple models, conducts multi-dimensional comprehensive analysis and evaluation of the tool, and generates a detailed inspection report. The report uses charts to display data and intuitively reflect the inspection results.

[0041] The human-computer interaction interface provides a user interface with a clear layout and reasonable arrangement of elements to improve user convenience. After the user performs the inspection task, the inspection results, including insulation performance, mechanical strength, and degree of wear, are displayed in a prominent position in the center of the interface. The interface has a history module, and users can query past records by time range and inspection item filtering. Inspection results that exceed preset thresholds and are abnormal will automatically generate warning information, which will be highlighted in eye-catching colors and flashing prompts to remind users. A powerful query engine is provided to support multiple query methods. Users can select inspection records to generate inspection reports in multiple formats with one click, allowing users to adjust the interface layout and color theme according to personal preferences. Built-in help documents and FAQs help users get started quickly and solve problems. There is a feedback portal where users can submit usage feedback and suggestions to optimize system functions and user experience.

[0042] Example 2:

[0043] The electric safety tool detection system and detection method of this embodiment particularly emphasize the application of a non-contact sensor array.

[0044] Infrared thermal imagers are used to detect the surface temperature distribution of tools, laser rangefinders are used to measure the deformation of tools, ultrasonic flaw detectors and matching probes are used to detect internal cracks and surface wear. The testing room maintains a constant temperature and humidity to reduce the impact of the external environment on infrared testing and ensure that the testing area is unobstructed. The laser rangefinder and ultrasonic flaw detector can accurately align with the testing point, clean the surface of the tool, remove dirt and coating that may affect the detection, and mark the predetermined measurement points and testing areas.

[0045] Through infrared thermal imaging to detect insulation performance and leakage current analysis, set the infrared thermal imager parameters, including temperature range and resolution, conduct a comprehensive scan of the tool, obtain the infrared radiation intensity of each point, convert the infrared radiation intensity data into temperature values, form a temperature distribution image of the tool surface, and calculate the average temperature T of the entire image. avg , the formula is: Among them, T ij is the temperature of the point with coordinates (i, j) in the image, N and M are the number of rows and columns of the image respectively. The standard deviation σ of the temperature is calculated as follows: A threshold of 3σ is set. If the difference from the average temperature exceeds the threshold, it is considered to be a temperature anomaly area. The temperature gradient of the temperature anomaly area is calculated using the formula: The heat generated by leakage current follows Joule's law Q = I 2 Rt, where Q is heat, I is current, R is resistance, the resistivity ρ is related to the geometry of the material, and t is time, through the heat conduction equation: Among them, α is the thermal diffusivity, c is the specific heat capacity, ρ is the material density, combined with thermal conductivity k, resistivity ρ, temperature gradient Parameters, analyze the possibility and approximate location of leakage current.

[0046] Use a laser rangefinder to test the mechanical strength. Aim the laser rangefinder at the predetermined measuring point of the tool and continuously measure the distance and record the data at a preset frequency of 10 times per second. Calculate the distance L by measuring the time difference Δt between the emitted light and the reflected light. The distance calculation formula is: Where c is the speed of light, which is approximately 299792458 m / s. A series of distance measurement data is obtained as L1, L2, L3, ..., L n , calculate the distance change between adjacent measurement points: ΔL i =L i+1 -L i , where i = 1, 2, 3, ..., n-1, calculate the average value of the distance change: Calculate the standard deviation of the distance variation: The distance L is calculated by analyzing the time series curve of the distance change and measuring the time difference Δt between the emitted light and the reflected light. The distance calculation formula is: Where c is the speed of light, which is approximately 299792458 m / s. A series of distance measurement data is obtained as L1, L2, L3, ..., L n , calculate the distance change between adjacent measurement points: ΔL i =L i+1 -L i , where i = 1, 2, 3, ..., n-1, calculate the average value of the distance change: Calculate the standard deviation of the distance variation: By analyzing the time series curve of distance change, observing the deformation trend of the tool during force and vibration, and combining the mechanical properties and geometric shape of the materials used in the tool, based on the relevant theories of material mechanics and structural mechanics, the reliability of the mechanical strength of the tool under the current force and vibration conditions is evaluated.

[0047] Wear and cracks are detected using an ultrasonic flaw detector. A probe that matches the tool material is selected and installed on the flaw detector. Multiple key inspection areas are selected on the tool surface. The flaw detector emits short ultrasonic pulses into the tool. The pulse propagation velocity v is determined by the material properties and is expressed as: Where E is the elastic modulus of the material, ρ is the material density, and the reflected signal is received and its waveform characteristics, including amplitude A and phase φ, are analyzed. The defect location L is calculated using the time-of-flight method using the following formula: Where t is the time difference between the transmitted pulse and the received reflected pulse. The waveform characteristics and probe beam diffusion characteristics are combined to identify the specific location, size, and shape of cracks and wear. The acoustic characteristics of the tool material and the usage history of the tool are combined to evaluate the degree of wear and its impact on tool performance.

[0048] Summarize all test data and form a detailed test report.

[0049] Example 3:

[0050] This embodiment provides an electric safety tool detection system and detection method thereof, focusing on a multi-dimensional analysis module.

[0051] The multi-dimensional analysis module receives data from the data acquisition and processing module, classifies and organizes the data according to different tool types, inspection items and time series, and forms a structured data set for classified storage.

[0052] Support vector machine (SVM) is selected for insulation status judgment, and the data related to the insulation performance of the tool are organized into feature vectors x i ∈R n , where n is the number of features and the corresponding label y i ∈{-1,1}, where -1 indicates good insulation and 1 indicates a problem. The Gaussian kernel function is selected as the kernel function of SVM, and its formula is: K(x i ,x j )=exp(-γ||x i -x j || 2 ), where γ is the parameter of the kernel function. The goal of SVM is to maximize the distance between the two types of data points and the hyperplane. The optimization problem is expressed as ξ i ≥0, where C is the penalty parameter, ξ i is the slack variable, φ(x i ) is a function that maps input data to a high-dimensional feature space, w is the normal vector of the hyperplane, and b is the intercept of the hyperplane. Using the Lagrange multiplier method, the above optimization problem is transformed into its dual problem: Solve the above dual problem and get the Lagrange multiplier α i The value of α is obtained according to the solution i , calculate w and b, For any support vector x j The final decision function is: For new tools insulation performance data x new , substitute it into the decision function to get the predicted label, so as to judge whether its insulation status is good or has problems.

[0053] Random forest is used to evaluate the strength level. The data related to the strength of the tools are organized into feature vectors and assigned strength level labels. Random sampling with replacement is performed from the original data set to create multiple bootstrap sample sets. For each bootstrap sample set, a decision tree is constructed. K features are randomly selected from the feature set. The best splitting features and splitting points are selected according to the Gini index criterion. The Gini index formula is: Where K is the number of categories, p k,t It is the proportion of samples belonging to category k in node t. Multiple decision trees are repeatedly constructed to form a random forest. For new tool data, it is input into each decision tree in the random forest for prediction to obtain prediction results of multiple intensity levels. Multiple prediction results are combined by voting to obtain the final intensity level evaluation.

[0054] Use convolutional neural network to analyze the wear condition. The labeled wear images are divided into training set, validation set and test set. The convolution kernel is used to perform convolution operation on the input image to extract features. The convolution operation is expressed as: Among them, Z ij is the element of the output feature map, W mn is the element of the convolution kernel, X i+m,j+n is the element of the input image, b is the bias term, the maximum pooling layer is applied to reduce the dimension of the feature map, the features are mapped to the output category through the fully connected layer, and the cross entropy loss function is used as the model loss function, which is expressed as: Where N is the number of samples, y i is the true label, is the prediction probability. The network parameters are optimized using the stochastic gradient descent algorithm. The new tool wear image is input to obtain the prediction result of the wear degree.

[0055] By integrating the output results of multiple models, we conduct a multi-dimensional analysis of tools and equipment. Based on the analysis and judgment results, we generate a detailed inspection report and use charts to display the data.

[0056] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

Claims

1. A power safety tool detection system, characterized in that: The system consists of the following components: Non-contact sensor array: Integrates infrared sensors, laser rangefinders, ultrasonic flaw detectors and other non-contact sensors to detect the insulation performance, mechanical strength and wear degree of tools; Data acquisition and processing module: responsible for receiving data collected by each sensor, performing preprocessing, denoising, filtering, and extracting key feature information; Multi-dimensional analysis module: Based on machine learning and deep learning algorithms, it conducts multi-dimensional analysis on the processed data, accurately determines the insulation status, strength level and wear of tools, and generates a test report; Human-computer interaction interface: Provides an intuitive operation interface to display test results, historical records and warning information, supports user query and report export functions, and improves user experience.

2. The power safety tool detection system according to claim 1, characterized in that: The non-contact sensor array detects insulation performance and leakage current analysis through infrared thermal imaging, sets infrared thermal imager parameters, performs infrared thermal imaging scanning on tools, obtains infrared radiation intensity at each point, converts the received infrared radiation intensity data into temperature values, forms a temperature distribution image on the tool surface, and calculates the average temperature of the entire image. The formula is: Among them, T ij is the temperature of the point with coordinates (i, j) in the image, N and M are the number of rows and columns of the image respectively, and the standard deviation σ of the temperature is calculated as follows: A threshold value of 3σ is set. If the difference from the average temperature exceeds the threshold, it is considered to be a temperature abnormality area. The temperature gradient of the temperature abnormality area is calculated, and the possibility and location of leakage current are evaluated by combining Joule's law and heat conduction equation.

3. The power safety tool detection system according to claim 1, characterized in that: The non-contact sensor array detects mechanical strength through a laser rangefinder and continuously measures the tool at a preset measurement frequency. The laser beam emitted by the laser rangefinder illuminates the surface of the tool and calculates the distance L by measuring the time difference Δt between the emitted light and the reflected light. The distance calculation formula is: Where c is the speed of light, and the distance change between adjacent measurement points is calculated: ΔL i =L i+1 -L i , calculate the average value of the distance change: Calculate the standard deviation of the distance change: By analyzing the time series curve of distance change, observing the deformation trend of the tool during force and vibration, and combining the mechanical properties and geometric shape of the materials used in the tool, the reliability of the mechanical strength of the tool under the current force and vibration conditions is evaluated.

4. The power safety tool detection system according to claim 1, characterized in that: The non-contact sensor array detects wear and cracks through an ultrasonic flaw detector. An ultrasonic flaw detector probe that matches the material of the tool is selected, and multiple detection areas are selected on the surface of the tool. The ultrasonic flaw detector emits a series of short ultrasonic pulses into the tool. The pulse propagation speed v is determined by the material properties and is expressed as: Where E is the elastic modulus of the material, ρ is the material density, the reflected ultrasonic signal is received, the received reflected signal is analyzed, its waveform characteristics are extracted, and the defect position L is calculated by the time flight method. The formula is: Where t is the time difference between the transmitted pulse and the received reflected pulse. Based on the amplitude change ΔA and phase change Δφ of the reflected signal and the beam diffusion characteristics of the probe, the location, size and shape of internal cracks and surface wear can be identified. The degree of wear can be evaluated based on the acoustic characteristics of the tool material and the usage history of the tool.

5. The power safety tool detection system according to claim 1, characterized in that: The data acquisition and processing module monitors and receives data streams from a variety of non-contact sensors in real time through a serial communication interface, converts the received raw data from a sensor-specific format into a unified data format, filters the data, applies a feature extraction algorithm to process the filtered data, extracts temperature anomaly areas, deformation degrees, and crack sizes, and stores the feature-extracted data in a database.

6. The power safety tool detection system according to claim 1, characterized in that: The multi-dimensional analysis module receives relevant data from the data acquisition and processing module, classifies and organizes the data according to different tool types, test items and time series to form a structured data set, selects a support vector machine to judge the insulation status, uses a random forest to evaluate the strength level, uses a convolutional neural network to analyze the wear condition, integrates the output results of multiple models, performs a multi-dimensional analysis of the tools, generates a detailed test report based on the results of the analysis and judgment, and uses charts to display the data.

7. The power safety tool detection system according to claim 6, characterized in that: The multi-dimensional analysis module selects the support vector machine to judge the insulation status, collects data related to the insulation performance of the tool, and represents it as a feature vector x i , the corresponding label y i ∈{-1,1}, -1 means good insulation, 1 means there is a problem, and the Gaussian kernel function is selected as the kernel function of the model, and its formula is: K(x i ,x j )=exp(-γ||x i -x j || 2 ), where γ is the parameter of the kernel function. The goal of SVM is to maximize the distance between the two types of data points and the hyperplane. The optimization problem is expressed as ξ i ≥0, where C is the penalty parameter, ξ i is the slack variable, φ(x i ) is a function that maps input data to a high-dimensional feature space, w is the normal vector of the hyperplane, b is the intercept of the hyperplane, and the Lagrange multiplier method is used to transform the above optimization problem into its dual problem and solve the obtained α i , calculate w and b, for any support vector x j The final decision function is: For new tools insulation performance data x new Substitute it into the decision function to get the predicted label, thereby judging whether its insulation status is good or has problems.

8. The electric power safety tool detection system according to claim 6, characterized in that: The multi-dimensional analysis module uses random forest to evaluate the intensity level, collects relevant data of tools and instruments, and represents it as a feature vector X, the corresponding intensity level label is Y, and performs random sampling with replacement from the original data set to create multiple self-service sample sets. For each self-service sample set, a decision tree is constructed, m features are randomly selected from the feature set, and the best splitting features and splitting points are selected according to the Gini index criterion. The Gini index formula is: Where K is the number of categories, p k,t is the proportion of samples belonging to category k in node t. Multiple decision trees are repeatedly constructed to form a random forest. For new tool data, it is input into each decision tree in the random forest to predict X new , obtain prediction results of multiple intensity levels, and integrate multiple prediction results by voting to obtain the final intensity level evaluation.

9. The power safety tool detection system according to claim 6, characterized in that: The multi-dimensional analysis module uses a convolutional neural network to analyze the wear condition, collects tool wear image data, marks the area and degree of wear, receives the wear image data, uses a convolution kernel to perform a convolution operation on the input image, and extracts features. The convolution operation is expressed as: Among them, Z ij is the element of the output feature map, W mn is the element of the convolution kernel, X i+m,j+n is the element of the input image, b is the bias term, the maximum pooling is used to reduce the dimension of the feature map, the extracted features are mapped to the final output category through the fully connected layer, and the cross entropy loss function is used as the model loss function, which is expressed as: Where N is the number of samples, y i is the true label, It is the prediction probability. The stochastic gradient descent algorithm is used to update the weights and biases of the network. The new tool image to be analyzed is input into the trained convolutional neural network to obtain the prediction result of the wear degree.

10. A method for detecting electric safety tools, characterized in that: The specific steps of this detection method are: Data collection and processing: Place the power safety tools to be tested in the detection area, start the multi-sensor fusion detection system, and each sensor synchronously collects multi-dimensional data of the tools. Use advanced image processing and signal analysis algorithms to pre-process, extract features and perform pattern recognition on the collected data. Multi-dimensional inspection and evaluation: Based on the processed data, comprehensive and multi-dimensional inspection and evaluation of tools are carried out, including: Insulation performance testing: Analyze the temperature distribution on the surface of the tool through infrared images, and combine laser scanning data to determine whether the insulation layer is damaged, aged or falling off; Mechanical strength assessment: Use laser sensors to accurately measure the size and shape changes of tools, and use ultrasonic waves to detect internal cracks and defects to comprehensively assess their mechanical strength; Wear degree detection: Use ultrasonic sensors to detect wear marks on the surface of tools and internal wear conditions to evaluate their service life; Automatic classification and identification: According to the test results, the tools are automatically classified into qualified, to be repaired and scrapped, and their test status and results are recorded; Data storage and report generation: The test data, results and classification information are stored in the database, allowing users to query historical records at any time. The system generates detailed test reports based on demand and supports export in multiple formats.

Images