High-voltage isolating switch equipment capable of intelligently detecting clamping pressure
By adopting a structural design using directional slotted copper tubes and low thermal expansion coefficient steel bars in high-voltage disconnect switches, combined with high-precision sensors and machine learning models, the accuracy and intelligence issues of clamping pressure monitoring have been solved. This enables high-precision real-time monitoring and intelligent early warning over a wide temperature range, improving the operational reliability and maintenance efficiency of the equipment.
Patent Information
- Application Number
- CN202510848603.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-24
- Publication Date
- 2025-11-14
AI Technical Summary
Existing high-voltage disconnector clamping pressure monitoring technology suffers from problems such as insufficient structural design and signal acquisition, significant impact from environmental factors, insufficient intelligent data processing, and inability to achieve high-precision pressure detection, resulting in low equipment reliability and maintenance efficiency.
By using directional slots inside copper tubes and installing steel bars with low thermal expansion coefficients, combined with high-precision sensors and machine learning-based predictive models, accurate monitoring and intelligent early warning of clamping pressure can be achieved.
High-precision, real-time clamping pressure monitoring is achieved over a wide temperature range, improving the intelligent operation and maintenance level of the equipment, enhancing the reliability and timeliness of fault early warning, and supporting predictive maintenance.
Smart Images
Figure CN120947872A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of high-voltage power equipment condition monitoring technology, and in particular to a high-voltage disconnecting switchgear with intelligent clamping pressure detection. Background Technology
[0002] High-voltage disconnect switches are critical components for the safe operation of power systems, and the clamping pressure between their contacts directly affects conductivity reliability and operational safety. However, existing monitoring technologies generally suffer from the following bottlenecks, severely restricting equipment reliability and maintenance efficiency:
[0003] 1. Insufficient structural design and signal acquisition: Traditional equipment is directly installed on the metal outer wall of the disconnector. Due to the high rigidity of the metal material itself, it is difficult to effectively collect small deformations, resulting in high noise in the monitoring data and making it difficult to make accurate predictions.
[0004] 2. Significantly affected by environmental factors: Environmental factors such as temperature and current magnitude cause drift in sensor signals, leading to increased fault warning errors, and traditional data processing algorithms struggle to achieve real-time intelligent compensation.
[0005] 3. Insufficient intelligence in data acquisition and processing systems: Existing systems generally adopt traditional signal processing and data recording methods, lacking adaptive compensation based on big data and machine learning, and are unable to extract key features from historical data and continuously optimize prediction models.
[0006] 4. Inability to achieve high-precision stress detection: Most existing monitoring systems remain at the level of simple recording of raw data and fixed threshold alarms, lacking deep data mining, complex pattern recognition, and adaptive learning capabilities. They cannot effectively extract hidden state evolution patterns from the massive historical data accumulated over long-term operation, making it difficult to achieve accurate early fault prediction and health status assessment.
[0007] Therefore, there is an urgent need for a new type of high-voltage disconnector pressure monitoring technology that can enhance the effective signal from the structural design, initially isolate thermal disturbances at the physical level, and use advanced intelligent algorithms for deep data fusion and accurate prediction, so as to completely solve the problem of high-precision measurement over a wide temperature range and improve the intelligent operation and maintenance level of the equipment. Summary of the Invention
[0008] This application provides a high-voltage disconnecting switch device with intelligent clamping pressure detection. Through collaborative innovation in structure, sensing, and algorithms, the device achieves accurate and reliable monitoring and intelligent early warning of the clamping pressure of the disconnecting switch over a wide temperature range.
[0009] To address the aforementioned technical problems, this application provides a high-voltage disconnecting switch device with intelligent clamping pressure detection, comprising: a copper tube, a steel bar, a sensor, and a data processing and prediction module; the copper tube has at least one directional slot for reducing local stiffness and guiding stress concentration; the steel bar is installed inside the copper tube and is fixedly connected to the inner wall of the copper tube; the steel bar is used to receive deformation signals transmitted from the directional slotted area of the copper tube, and the thermal expansion coefficient of the steel bar material is lower than that of the copper tube material of the disconnecting switch, so as to achieve thermal isolation between the sensor installation position and the copper tube; the sensor is fixedly installed on the steel bar; the sensor includes a full-bridge strain gauge for detecting micro-strain of the steel bar and a temperature sensor for detecting the local temperature of the installation point of the steel bar; the data processing and prediction module is used to receive and preprocess the strain signal from the full-bridge strain gauge and the temperature signal from the temperature sensor, construct a feature vector containing the interaction relationship between the strain signal and the temperature signal for temperature compensation processing, use a machine learning-based prediction model to predict the actual clamping pressure value borne by the disconnecting switch in real time according to the constructed feature vector, and generate an early warning signal when the predicted pressure value exceeds a preset safety threshold.
[0010] In some exemplary embodiments, two symmetrical directional slots are formed on the copper tube, and two parallel steel bars are installed inside the copper tube; the parallel planes of the two steel bars are perpendicular to the parallel planes formed by the two directional slots.
[0011] In some exemplary embodiments, the directional slot has a length of 400 mm and a width of 5 mm.
[0012] In some exemplary embodiments, the steel bar is fixed inside the copper tube by a preset positioning hole and a high-temperature resistant fixing screw.
[0013] In some exemplary embodiments, both the full-bridge strain gauges and the temperature sensors are multiple, and the full-bridge strain gauges and the temperature sensors are distributed on the steel bar in an alternating or grouped manner.
[0014] In some exemplary embodiments, multiple sets of sensors are distributed and bonded in an alternating or grouped manner on the inner surface of each steel bar, along its length, in the key force transmission area.
[0015] In some exemplary embodiments, the machine learning-based prediction model includes a combination of graph convolutional networks and attention mechanisms.
[0016] In some exemplary embodiments, constructing a feature vector containing the interaction relationship between the strain signal and the temperature signal includes: calculating the product term of the strain signal and the temperature signal, as shown in the following equation:
[0017] FT=(F*T) / (F 2 +T 2 )
[0018] Where F represents the strain signal and T represents the temperature signal.
[0019] In some exemplary embodiments, the copper tube has a length of 800 mm, an outer diameter of 40 mm, and an inner diameter of 30 mm.
[0020] In some exemplary embodiments, the steel bar is a stainless steel bar with a length of 760 mm, a width of 15 mm, and a thickness of 10 mm.
[0021] The technical solution provided in this application has at least the following advantages:
[0022] This application provides a high-voltage disconnecting switchgear with intelligent clamping pressure detection, comprising: a copper tube, a steel bar, a sensor, and a data processing and prediction module; the copper tube has at least one directional slot for reducing local stiffness and guiding stress concentration; the steel bar is installed inside the copper tube and is fixedly connected to the inner wall of the copper tube; the steel bar is used to receive deformation signals transmitted by the directional slotted area of the copper tube, and the thermal expansion coefficient of the steel bar material is lower than that of the copper tube material of the disconnecting switch, so as to achieve thermal isolation between the sensor installation position and the copper tube; the sensor is fixedly installed on the steel bar; the sensor includes a full-bridge strain gauge for detecting micro-strain of the steel bar and a temperature sensor for detecting the local temperature of the steel bar installation point; the data processing and prediction module is used to receive and preprocess the strain signal from the full-bridge strain gauge and the temperature signal from the temperature sensor, construct a feature vector containing the interaction relationship between the strain signal and the temperature signal for temperature compensation processing, and use a machine learning-based prediction model to predict the actual clamping pressure value borne by the disconnecting switch in real time according to the constructed feature vector, and generate an early warning signal when the predicted pressure value exceeds a preset safety threshold.
[0023] This high-voltage disconnecting switchgear achieves accurate and reliable monitoring and intelligent early warning of the disconnecting switch clamping pressure over a wide temperature range through collaborative innovation in structure, sensing, and algorithms. Its technical solution includes the following main aspects:
[0024] 1. Structural reinforcement and signal decoupling design
[0025] This application innovates the structure at the physical level by using a highly conductive alloy copper tube of specific dimensions (800mm long, 40mm outer diameter, and 30mm inner diameter) as the main body. A symmetrical directional slot (400mm long and 5mm wide) is precisely machined in the central area of the tube. This significantly reduces local stiffness to concentrate and initially amplify the deformation signal caused by pressure. At the same time, two 304 stainless steel bars (760mm long, 15mm wide, and 10mm thick) are installed in parallel inside the copper tube through pre-set positioning holes and fixing screws. Their high stiffness and low coefficient of thermal expansion (compared to the copper tube) not only make them a reliable sensing platform to further transmit and even amplify the deformation signal (the parallel planes of the two steel bars are perpendicular to the parallel plane of the slot), but more importantly, they effectively physically isolate the temperature-sensitive sensor from the main heat source (copper tube). This enhances the target signal at the source and suppresses thermal noise interference, laying a solid foundation for subsequent high-precision measurements.
[0026] 2. High-precision multi-physical quantity sensing strategy
[0027] To accurately capture the weak signals after structural enhancement, this application employs a high-precision multi-physical quantity sensing strategy. Five sets of high-sensitivity full-bridge strain gauges and high-precision temperature sensors are precisely and alternately arranged along the length of the surface of the key stress-transfer area of the internal stainless steel strip, enabling synchronous and accurate measurement of mechanical strain and corresponding local temperature. Combined with a high-performance front-end signal conditioning module including a low-noise instrumentation amplifier, high-order filters, and a 16-bit high-resolution analog-to-digital converter (ADC), high fidelity is ensured during the conversion from physical to digital signals, providing high-quality raw data input for subsequent intelligent algorithms.
[0028] 3. AI-driven intelligent data fusion and prediction engine
[0029] The core intelligence of this application lies in its AI-driven data processing and prediction engine. This engine first employs a local outlier measurement algorithm to denoise and clean the raw data, then performs crucial deep feature engineering to construct interactive features that reflect the complex nonlinear coupling relationship between strain (F) and temperature (T). Utilizing these high-dimensional features processed by Min-Max normalization, and combining them with a powerful spatiotemporal fusion machine learning model—that is, combining a graph convolutional network (GCN) with a cross-attention mechanism to process local information from the past 20 time points and 20 global similarity pattern information—the system can overcome the limitations of traditional compensation methods and achieve high-precision, real-time prediction of the disconnector switch clamping pressure of 50 kN over a wide temperature range of -20°C to 150°C, thus realizing real-time detection of the disconnector switch pressure.
[0030] 4. System Integration and Intelligent Platform
[0031] To maximize the practical application and value of the technology, this application constructs a complete system integration and intelligent platform. It employs reliable low-power long-range wireless communication technology (LoRa) and supports edge computing, ensuring flexible and secure data transmission and initial processing in complex environments. The backend is built on a powerful multi-terminal monitoring and decision support platform based on modern Web technology. Through a user-friendly interface, it provides functions such as real-time stress curves, historical data visualization, intelligent alarms, health assessment reports, and predictive maintenance suggestions, transforming precise monitoring capabilities into efficient and intelligent equipment operation and maintenance management tools. Attached Figure Description
[0032] One or more embodiments are illustrated by way of example with reference to the accompanying drawings. These illustrations do not constitute a limitation on the embodiments, and unless otherwise stated, the figures in the drawings are not to be limited by scale.
[0033] Figure 1 This is a schematic diagram of the structure of a high-voltage disconnecting switchgear with intelligent clamping pressure detection provided in an embodiment of this application.
[0034] Figure 2 This is a diagram of the copper tube and directional slotted structure provided in this application.
[0035] Figure 3 This is the installation and mechanical transfer diagram of the stainless steel strip provided in this application.
[0036] Figure 4 This is a schematic diagram of the sensor arrangement and front-end signal conditioning provided in this application.
[0037] Figure 5 This is a schematic diagram of the machine learning prediction model for the device provided in this application. Detailed Implementation
[0038] As can be seen from the background technology, existing monitoring technologies generally have many bottlenecks, which seriously restrict the reliability of equipment operation and maintenance efficiency. Therefore, there is an urgent need for a new type of high-voltage disconnector pressure monitoring technology that can enhance the effective signal from the structural design, initially isolate thermal disturbances at the physical level, and use advanced intelligent algorithms for deep data fusion and accurate prediction, so as to completely solve the problem of high-precision measurement over a wide temperature range and improve the intelligent operation and maintenance level of the equipment.
[0039] To address the aforementioned technical problems, this application provides a high-voltage disconnecting switchgear with intelligent clamping pressure detection, comprising: a copper tube, a steel bar, a sensor, and a data processing and prediction module; the copper tube has at least one directional slot for reducing local stiffness and guiding stress concentration; the steel bar is installed inside the copper tube and is fixedly connected to the inner wall of the copper tube; the steel bar is used to receive deformation signals transmitted from the directional slotted area of the copper tube, and the thermal expansion coefficient of the steel bar material is lower than that of the copper tube material of the disconnecting switch, so as to achieve thermal isolation between the sensor installation position and the copper tube; the sensor is fixedly installed on the steel bar; the sensor includes a full-bridge strain gauge for detecting micro-strain of the steel bar and a temperature sensor for detecting the local temperature of the steel bar installation point; the data processing and prediction module is used to receive and preprocess the strain signal from the full-bridge strain gauge and the temperature signal from the temperature sensor, construct a feature vector containing the interaction relationship between the strain signal and the temperature signal for temperature compensation processing, and use a machine learning-based prediction model to predict the actual clamping pressure value borne by the disconnecting switch in real time according to the constructed feature vector, and generate an early warning signal when the predicted pressure value exceeds a preset safety threshold.
[0040] The embodiments of this application will now be described in detail with reference to the accompanying drawings. However, those skilled in the art will understand that many technical details have been provided in the embodiments of this application to facilitate a better understanding of the application. However, the technical solutions claimed in this application can be implemented even without these technical details and various variations and modifications based on the following embodiments.
[0041] See Figure 1 This application provides a high-voltage disconnecting switchgear with intelligent clamping pressure detection, comprising: a copper tube, a steel bar, a sensor, and a data processing and prediction module; the copper tube has at least one directional slot in its predetermined clamping force area to reduce local stiffness and guide stress concentration; the steel bar is installed inside the copper tube and is fixedly connected to the inner wall of the copper tube; the steel bar is used to receive deformation signals transmitted by the directional slot area of the copper tube, and the thermal expansion coefficient of the steel bar material is lower than that of the copper tube material of the disconnecting switch, so as to achieve thermal isolation between the sensor installation position and the copper tube; the sensor is fixedly installed on the steel bar; the sensor includes a full-bridge strain gauge for detecting micro-strain of the steel bar and a temperature sensor for detecting the local temperature of the steel bar installation point; the data processing and prediction module is used to receive and preprocess the strain signal from the full-bridge strain gauge and the temperature signal from the temperature sensor, construct a feature vector containing the interaction relationship between the strain signal and the temperature signal for temperature compensation processing, use a machine learning-based prediction model to predict the actual clamping pressure value borne by the disconnecting switch in real time according to the constructed feature vector, and generate an early warning signal when the predicted pressure value exceeds a preset safety threshold.
[0042] In some embodiments, two symmetrical directional slots are formed on the copper tube, and two parallel steel bars are installed inside the copper tube; the parallel planes of the two steel bars are perpendicular to the parallel planes formed by the two directional slots.
[0043] In some embodiments, the directional slot has a length of 400 mm and a width of 5 mm.
[0044] In some embodiments, the copper tube has a length of 800 mm, an outer diameter of 40 mm, and an inner diameter of 30 mm.
[0045] Figure 2 A diagram of a copper tube with a directional slotted structure is shown. Figure 2 The dimensions of the copper tube (length gc = 800 mm, outer diameter wj = 40 mm, inner diameter nj = 30 mm) and the cross-section and precision machining details of the groove with length cc = 400 mm and width ck = 5 mm processed in the predetermined area are shown, illustrating the effect of the groove geometry on reducing local rigidity in the overall stress.
[0046] In some embodiments, the steel strip is fixed inside the copper tube through a pre-set positioning hole and a high-temperature resistant fixing screw.
[0047] In some embodiments, the steel bar is a stainless steel bar with a length of 760 mm, a width of 15 mm, and a thickness of 10 mm.
[0048] Figure 3 The diagram illustrates the installation and mechanical transfer of stainless steel bars. From... Figure 3 It can be seen that the stainless steel bar is firmly installed inside the copper tube using the positioning holes and fixing screws on the inner wall, and the mechanical principle of the bar in transmitting, amplifying and stabilizing the overall force signal can be observed.
[0049] Figure 4 A schematic diagram of sensor arrangement and front-end signal conditioning is shown. Figure 4 The diagram shows the alternating distribution of full-bridge strain gauges, temperature sensors, and other environmental sensors on a stainless steel bar.
[0050] In some embodiments, both full-bridge strain gauges and temperature sensors are multiple, and the full-bridge strain gauges and temperature sensors are distributed on the steel strip in an alternating or grouped manner.
[0051] In some embodiments, multiple sets of sensors are distributed and bonded in an alternating or grouped manner on the inner surface of each steel bar, along its length, in the key force transmission area.
[0052] In some embodiments, the machine learning-based prediction model includes a combination of graph convolutional networks and attention mechanisms.
[0053] Figure 5The device's machine learning prediction model is shown. Figure 5 It demonstrates the complete process from data preprocessing, feature extraction, machine learning model training, online adaptive updates, and fault warning generation, and clarifies the data transmission and feedback paths between modules.
[0054] In some embodiments, constructing a feature vector containing the interaction between the strain signal and the temperature signal includes: calculating the product term of the strain signal and the temperature signal, as shown in the following equation:
[0055] FT=(F*T) / (F 2 +T 2 )
[0056] Where F represents the strain signal and T represents the temperature signal.
[0057] This high-voltage disconnecting switchgear achieves accurate and reliable monitoring and intelligent early warning of the disconnecting switch clamping pressure over a wide temperature range through collaborative innovation in structure, sensing, and algorithms. Its technical solution includes the following main aspects:
[0058] 1. Structural reinforcement and signal decoupling design
[0059] This application innovates the structure at the physical level by using a highly conductive alloy copper tube of specific dimensions (800mm long, 40mm outer diameter, and 30mm inner diameter) as the main body. A symmetrical directional slot (400mm long and 5mm wide) is precisely machined in the central area of the tube. This significantly reduces local stiffness to concentrate and initially amplify the deformation signal caused by pressure. At the same time, two 304 stainless steel bars (760mm long, 15mm wide, and 10mm thick) are installed in parallel inside the copper tube through pre-set positioning holes and fixing screws. Their high stiffness and low coefficient of thermal expansion (compared to the copper tube) not only make them a reliable sensing platform to further transmit and even amplify the deformation signal (the parallel planes of the two steel bars are perpendicular to the parallel plane of the slot), but more importantly, they effectively physically isolate the temperature-sensitive sensor from the main heat source (copper tube). This enhances the target signal at the source and suppresses thermal noise interference, laying a solid foundation for subsequent high-precision measurements.
[0060] 2. High-precision multi-physical quantity sensing strategy
[0061] To accurately capture the weak signals after structural enhancement, this application employs a high-precision multi-physical quantity sensing strategy. Five sets of high-sensitivity full-bridge strain gauges and high-precision temperature sensors are precisely and alternately arranged along the length of the surface of the key stress-transfer area of the internal stainless steel strip, enabling synchronous and accurate measurement of mechanical strain and corresponding local temperature. Combined with a high-performance front-end signal conditioning module including a low-noise instrumentation amplifier, high-order filters, and a 16-bit high-resolution analog-to-digital converter (ADC), high fidelity is ensured during the conversion from physical to digital signals, providing high-quality raw data input for subsequent intelligent algorithms.
[0062] 3. AI-driven intelligent data fusion and prediction engine
[0063] The core intelligence of this application lies in its AI-driven data processing and prediction engine. This engine first employs a local outlier measurement algorithm to denoise and clean the raw data, then performs crucial deep feature engineering to construct interactive features that reflect the complex nonlinear coupling relationship between strain (F) and temperature (T). Utilizing these high-dimensional features processed by Min-Max normalization, and combining them with a powerful spatiotemporal fusion machine learning model—that is, combining a graph convolutional network (GCN) with a cross-attention mechanism to process local information from the past 20 time points and 20 global similarity pattern information—the system can overcome the limitations of traditional compensation methods and achieve high-precision, real-time prediction of the disconnector switch clamping pressure of 50 kN over a wide temperature range of -20°C to 150°C, thus realizing real-time detection of the disconnector switch pressure.
[0064] 4. System Integration and Intelligent Platform
[0065] To maximize the practical application and value of the technology, this application constructs a complete system integration and intelligent platform. It employs reliable low-power long-range wireless communication technology (LoRa) and supports edge computing, ensuring flexible and secure data transmission and initial processing in complex environments. The backend is built on a powerful multi-terminal monitoring and decision support platform based on modern Web technology. Through a user-friendly interface, it provides functions such as real-time stress curves, historical data visualization, intelligent alarms, health assessment reports, and predictive maintenance suggestions, transforming precise monitoring capabilities into efficient and intelligent equipment operation and maintenance management tools.
[0066] The high-voltage disconnect switchgear with intelligent clamping pressure detection provided in this application will be described in detail below through specific embodiments.
[0067] The main body and structural assembly of the disconnecting switch include:
[0068] (1) Machining and slotting of copper tubes for disconnecting switch body
[0069] High-conductivity, high-strength alloy copper material was selected and machined into a main copper tube of the designed dimensions (tube length gc = 800 mm, outer diameter wj = 40 mm, inner diameter nj = 30 mm). Using a high-precision CNC milling machine, two symmetrical directional slots were machined symmetrically in the central region of the copper tube, perpendicular to the radial direction (i.e., the expected main stress direction). Each slot has a length cc = 400 mm and a width ck = 5 mm (see...). Figure 2 The purpose of slotting is to reduce the local bending stiffness of the area, so that when the disconnecting switch contacts close and apply a clamping force F, the area will produce more significant radial deformation than when it is not slotted, thereby amplifying the mechanical signal related to the pressure F.
[0070] (2) Installation of stainless steel strips
[0071] Two precisely sized steel bars (760mm long, 15mm wide, and 10mm thick) were machined from 304 stainless steel, a material known for its high elastic modulus and low coefficient of thermal expansion. Precision laser-guided drilling was used to drill holes at predetermined positions on the inner wall of the copper tube. The two stainless steel bars were then installed parallel to each other inside the copper tube (see...). Figure 3 The steel bars are arranged so that their length is parallel to the axis of the copper tube, and their width (i.e., the parallel plane) is perpendicular to the parallel plane formed by the two slots on the copper tube. The steel bars are secured to the positioning holes on the inner wall of the copper tube with high-temperature resistant, vibration-resistant stainless steel fixing screws, ensuring a firm connection and effective force transmission during long-term operation. The steel bars not only serve as mounting bases for subsequent sensors, but their high rigidity also helps to transmit and potentially amplify the deformation transmitted from the copper tube. Furthermore, their thermal expansion difference relative to the copper tube and their internal position help to isolate some thermal interference.
[0072] (3) Sensor network construction and front-end data processing
[0073] On the inner surface of each stainless steel bar, along its length, in key stress-transfer areas, multiple sets of sensors are alternately and distributedly bonded using high-temperature resistant, high-insulation industrial-grade epoxy resin adhesive. Each set of sensors includes a high-sensitivity full-bridge strain gauge and a high-precision temperature sensor adjacent to it (see...). Figure 4A full-bridge strain gauge maximizes the compensation for the temperature effect on the resistance wire itself and improves the sensitivity of strain measurement. Each sensor is connected to a front-end signal conditioning circuit module. This module, typically integrated on a PCB, includes: a precision regulated power supply for the strain gauge bridge, a low-noise instrumentation amplifier for amplifying weak strain signals, a filter for filtering high-frequency interference, and a converter for converting the amplified and filtered analog signal into a 16-bit high-resolution digital signal. The temperature sensor signal is also processed by corresponding conditioning circuitry. All digitized sensor data (F1, T1, F2, T2, F3, T3, F4, T4, F5, T5) is transmitted to the data processing unit via an internal bus.
[0074] (4) Intelligent data acquisition, processing and prediction
[0075] Data from each node is collected in real time, and noise is removed by the preprocessing module. Then, feature engineering is used to construct temperature compensation features, which are normalized and input into the graph convolutional machine learning model to achieve feature extraction. The model parameters are updated in real time through backpropagation using the difference between the predicted pressure value and the actual pressure value.
[0076] The specific steps for denoising and removing outliers from the sampled data are as follows:
[0077] 1) For a data point p, its k-distance (k-dist) is defined as the distance from p to its k-th nearest neighbor as k-dist(p) = dist(p, o k ), where o k It is the k-th nearest neighbor of p, where dist(p,o) k ) is p and o k The Euclidean distance between them.
[0078] 2) The reachability distance is calculated as follows: for a data point p and its neighbor o, the reachability distance is defined as reach-dist. k (p,o)=max{dist(p,o),k-dist(o)}.
[0079] 3) Calculate Localized To Density (LTD). LTD measures the local density of a data point p and is defined as follows: Where N k (p) is the set of k-nearest neighbors of p.
[0080] 4) Assess the degree of anomaly by comparing the local density of data point p with the local density of its neighbors: If C k If (p) is greater than 1, it means that the local density of p is lower than that of its neighbors, and an outlier can be defined, i.e., an outlier value.
[0081] 5) The equipment's sensors include 5 full-bridge strain gauges and 5 temperature sensors, with corresponding data acquisition data (F1, T1, F2, T2, F3, T3, F4, T4, F5, T5), where T1 corresponds to the temperature sensor for F1, and the same applies to the other sensors. The degree of anomaly in the sampled data is assessed by calculating the local density of each sampled data point and its neighbors. When sampled data is detected as anomaly, that data is deleted; similarly, all anomaly data is deleted.
[0082] The temperature compensation feature is constructed as follows: The equipment's sensors include 5 full-bridge strain gauges and 5 temperature sensors, with corresponding data acquisition values (F1, T1, F2, T2, F3, T3, F4, T4, F5, T5), where T1 corresponds to the temperature sensor F1, and the same applies to the other sensors. To construct a temperature compensation feature and eliminate the influence of temperature on the equipment's pressure value detection, a new feature, FT, is constructed. i =(F i *T i ) / (F i 2 +T i 2 ), where i = 1, 2, 3, 4, 5. FT i =(F i *T i ) / (F i 2 +T i 2 By fusing F i and T i This creates second-order properties and enhances the nonlinear relationship. Therefore, the final eigenvector at time t is F. t =(F1,T1,FT1,F2,T2,FT2,F3,T3,FT3,F4,T4,FT4,F5,T5,FT5).
[0083] Data normalization: To enhance the robustness of the prediction model, the feature vectors need to be normalized. The normalization process is min-max normalization, and its specific steps are as follows: x i Let x be the i-th eigenvector, x be the i-th column vector of all eigenvectors, max(x) be the maximum value of the i-th column vector of the eigenvectors, and min(x) be the minimum value of the i-th column vector of the eigenvectors. After data normalization, all features will be normalized to 0 to 1.
[0084] Network Model Learning: This application will sample a neural network model to predict equipment pressure values. The model input is... and in The matrix constructed from the combined vectors from the current time t-20 to time t represents the local information at time t, and its corresponding dimension is (20, 15). The matrix formed by the 20 most similar eigenvectors between the current time t and the eigenvectors at previous times represents the global information at time t, with dimensions (20, 15). This is achieved through calculation... The similarity of vectors at each time step is used to construct a graph-structured data. Similarly, calculate The similarity of vectors at each time step is used to construct a graph-structured data. After two layers of graph convolutional network After two layers of graph convolutional network Where x1 is The feature vector of the middle node for The adjacency matrix, W1 (0) W1 represents the network weights of the first layer of the first-path graph convolution. (1) The network weights for the first-path, second-layer graph convolution are given by H1, where H1 is the output of the first-path graph convolution. Similarly, x2 is... The feature vector of the middle node for The adjacency matrix, The network weights for the second path, first layer graph convolution. The network weights for the second-path, second-layer graph convolution are given, and H2 is the output of the second-path graph convolution. Then, H1 and H2 are input into a cross-attention module for fusing the output vectors of the two graph convolutions. The cross-attention module is... Among them W Q W K W V The corresponding weights are d, and d = 128 is the scaling factor. Finally, The input is expanded and fed into a fully connected network, and the final output is given. The model's loss function is... Where y i Actual pressure value at time t, p i Where N represents the predicted stress value, and N is the number of samples. This application sets the learning rate to 0.0001 and uses the Adma optimization method to learn the weights in the model.
[0085] The model prediction results of the equipment in this application are shown in Table 1, taking the actual pressure of 50 as an example.
[0086] Table 1. Differences between 50 KGN pressure values and predicted pressure values when the temperature changes from -20℃ to 150℃.
[0087]
[0088]
[0089] Similarly, when the copper tube temperature is 100℃, the comparison between the equipment's predicted and actual values is shown in Table 2 as the pressure changes from 0 to 150KGN.
[0090] Table 2 Comparison of predicted and actual values of equipment when the pressure changes from 0 to 150 KGN at a copper tube temperature of 100℃.
[0091]
[0092] To further verify the accuracy of the pressure prediction value of the device in this application, this application adopts a random sampling method, sampling different pressures and temperatures, observing the device temperature and predicted pressure values, and displaying them in a scatter plot as shown in Table 3.
[0093] Table 3 shows the differences between predicted and actual values of temperature and pressure at different sampling pressures and temperatures.
[0094]
[0095]
[0096] (5) Wireless communication, cloud scheduling and remote control
[0097] The device employs a dynamic early warning mechanism for its safety threshold settings. When predicted data exceeds this threshold, the central unit triggers an alarm, records detailed data, and uploads it to local storage and a cloud platform for comprehensive monitoring. The device utilizes a multi-mode communication solution supporting LoRa, ensuring stable communication even in complex network environments. Furthermore, it incorporates a dual-channel encryption module to ensure data transmission complies with international security standards.
[0098] In summary, the high-voltage disconnecting switchgear with intelligent clamping pressure detection provided in this application has the following advantages:
[0099] (1) High-precision pressure monitoring over a wide temperature range has been achieved.
[0100] By using an innovative "grooved copper tube + internal steel bar" structure to enhance the sensing design, signal amplification and thermal disturbance isolation are initially achieved at the physical level. Combined with an intelligent temperature compensation algorithm based on deep feature engineering and advanced AI models, the force-thermal coupling effect is accurately decoupled at the information level. This fundamentally solves the core pain point of low pressure measurement accuracy and poor consistency in the existing technology over a wide temperature range (-20℃ to 150℃), as shown in the example data, with high prediction accuracy.
[0101] (2) Significantly improved the reliability and timeliness of fault warning.
[0102] High-precision real-time pressure prediction can more sensitively and accurately reflect minute abnormal changes in clamping force. Combined with dynamic early warning thresholds, it can achieve earlier and more reliable fault warnings, effectively preventing equipment failures or safety accidents caused by abnormal pressure.
[0103] (3) Enable predictive maintenance and optimize operation and maintenance strategies.
[0104] Continuous and accurate pressure status monitoring data, combined with the health assessment and trend prediction capabilities of AI models, makes it possible to shift from traditional scheduled maintenance or post-failure repair to predictive maintenance based on actual conditions. Maintenance personnel can precisely schedule maintenance plans based on the health indices and maintenance recommendations provided by the system, avoiding unnecessary downtime for repairs, reducing total maintenance costs, and improving equipment asset utilization.
[0105] (4) The system has high overall robustness and intelligence.
[0106] The integrated hardware and software design, high signal-to-noise ratio sensing front-end, interference-resistant wireless communication, powerful AI data analysis engine, and feature-rich, user-friendly monitoring platform enable the entire monitoring system to have high reliability, high automation, high intelligence, and good ease of use in complex industrial environments.
[0107] This application provides a high-voltage disconnecting switchgear with intelligent clamping pressure detection, comprising: a copper tube, a steel bar, a sensor, and a data processing and prediction module; the copper tube has at least one directional slot in its predetermined clamping force area to reduce local stiffness and guide stress concentration; the steel bar is installed inside the copper tube and is fixedly connected to the inner wall of the copper tube; the steel bar is used to receive deformation signals transmitted by the directional slot area of the copper tube, and the thermal expansion coefficient of the steel bar material is lower than that of the copper tube material of the disconnecting switch, so as to achieve thermal isolation between the sensor installation position and the copper tube; the sensor is fixedly installed on the steel bar; the sensor includes a full-bridge strain gauge for detecting micro-strain of the steel bar and a temperature sensor for detecting the local temperature of the steel bar installation point; the data processing and prediction module is used to receive and preprocess the strain signal from the full-bridge strain gauge and the temperature signal from the temperature sensor, construct a feature vector containing the interaction relationship between the strain signal and the temperature signal for temperature compensation processing, use a machine learning-based prediction model to predict the actual clamping pressure value borne by the disconnecting switch in real time according to the constructed feature vector, and generate an early warning signal when the predicted pressure value exceeds a preset safety threshold. This application aims to break through the limitations of traditional monitoring methods by combining physical-level signal enhancement with information-level intelligent processing to achieve full-temperature-range, high-precision, real-time online monitoring and fault early warning of disconnector clamping pressure, and to provide key technical support for realizing full life-cycle health management and predictive maintenance of equipment.
[0108] Those skilled in the art will understand that the above-described embodiments are specific examples of implementing this application, and in practical applications, various changes in form and detail may be made without departing from the spirit and scope of this application. Any person skilled in the art can make their own modifications and alterations without departing from the spirit and scope of this application; therefore, the scope of protection of this application should be determined by the scope defined in the claims.
Claims
1. A high-voltage disconnector device with intelligent clamping pressure detection, characterized in that, include: Copper pipes, steel bars, sensors, and data processing and prediction modules; The copper tube is provided with at least one directional slot for reducing local stiffness and guiding stress concentration. The steel bar is installed inside the copper tube and is fixedly connected to the inner wall of the copper tube; the steel bar is used to receive the deformation signal transmitted by the directional slotted area of the copper tube, and the thermal expansion coefficient of the steel bar material is lower than that of the material of the disconnecting switch copper tube, so as to achieve thermal isolation between the sensor installation position and the copper tube; The sensor is fixedly mounted on the steel bar; the sensor includes a full-bridge strain gauge for detecting micro-strain of the steel bar and a temperature sensor for detecting the local temperature at the mounting point of the steel bar; The data processing and prediction module is used to receive and preprocess the strain signal from the full-bridge strain gauge and the temperature signal from the temperature sensor, construct a feature vector containing the interaction relationship between the strain signal and the temperature signal for temperature compensation processing, use a machine learning-based prediction model to predict the actual clamping pressure value of the disconnecting switch in real time according to the constructed feature vector, and generate an early warning signal when the predicted pressure value exceeds a preset safety threshold.
2. The high-voltage disconnector switchgear with intelligent clamping pressure detection according to claim 1, characterized in that, The copper tube has two symmetrical directional slots, and two parallel steel bars are installed inside the copper tube; the parallel plane of the two steel bars is perpendicular to the parallel plane formed by the two directional slots.
3. The high-voltage disconnecting switchgear with intelligent clamping pressure detection according to claim 2, characterized in that, The directional slot has a length of 400mm and a width of 5mm.
4. The high-voltage disconnector switchgear with intelligent clamping pressure detection according to claim 1, characterized in that, The steel bar is fixed inside the copper tube through a pre-set positioning hole and a high-temperature resistant fixing screw.
5. The high-voltage disconnector switchgear with intelligent clamping pressure detection according to claim 1, characterized in that, Both the full-bridge strain gauge and the temperature sensor comprise multiple units, and the full-bridge strain gauge and the temperature sensor are distributed on the steel bar in an alternating or grouped manner.
6. The high-voltage disconnector switchgear with intelligent clamping pressure detection according to claim 5, characterized in that, On the inner surface of each steel bar, in the key force transmission area along its length, multiple sets of sensors are distributed and bonded in an alternating or grouped manner using epoxy resin adhesive.
7. The high-voltage disconnecting switchgear with intelligent clamping pressure detection according to claim 1, characterized in that, The machine learning-based prediction model includes a combination of graph convolutional networks and attention mechanisms.
8. The high-voltage disconnecting switchgear with intelligent clamping pressure detection according to claim 1, characterized in that, Constructing a feature vector containing the interaction relationship between the strain signal and the temperature signal includes: The product term of the strain signal and the temperature signal is calculated as follows: FT=(F*T) / (F 2 +T 2 ) Where F represents the strain signal and T represents the temperature signal.
9. The high-voltage disconnector switchgear with intelligent clamping pressure detection according to claim 1, characterized in that, The copper tube is 800mm long, 40mm in outer diameter, and 30mm in inner diameter.
10. The high-voltage disconnecting switchgear with intelligent clamping pressure detection according to claim 1, characterized in that, The steel bar is a stainless steel bar, with a length of 760mm, a width of 15mm, and a thickness of 10mm.