Arc extinguishing system, independent three-phase switch, and method for predicting electrical life of independent three-phase switch
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SOUTHELEC CO LTD
- Filing Date
- 2025-12-23
- Publication Date
- 2026-06-02
Smart Images
Figure CN121394206B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of electrical switchgear and its condition monitoring and fault prediction technology, specifically to arc extinguishing systems, independent three-phase switches, and methods for predicting the electrical life of independent three-phase switches. Background Technology
[0002] Existing independent three-phase switches have significant flaws in the connection method between the top and middle covers. While traditional bolt connections facilitate disassembly, they are prone to misalignment of the connectors during actual use, leading to decreased connection stability. This misalignment not only affects the overall structural strength of the switch but may also cause misalignment of internal components, thereby impacting the switch's operational performance and lifespan.
[0003] In terms of arc extinguishing, existing technologies mainly rely on a single arc-extinguishing chamber structure to eliminate the arc. However, during the tripping process of moving and stationary contacts, the arc's movement path is multidirectional. In addition to the arc moving towards the arc-extinguishing chamber, a considerable portion of the arc will transfer and spread along the terminal block. If this arc spreading along the terminal block is not eliminated in time, it will not only accelerate the burning of the terminal block but may also cause safety hazards. Existing technologies lack effective means to eliminate this arc spreading along the terminal block, resulting in unsatisfactory arc extinguishing effects.
[0004] Meanwhile, independent three-phase switches are key protection components in power distribution systems. Their electrical life is mainly limited by the performance degradation of the arc-extinguishing system. Traditional life prediction methods are mostly based on simple operation counts or cumulative breaking current (I²t) calculations. These methods have significant shortcomings: firstly, they treat the switch as a complete black box, failing to reflect the true degradation status of the complex internal arc-extinguishing structure; existing technologies urgently need improvement to address these issues. Summary of the Invention
[0005] The purpose of this invention is to overcome the shortcomings of existing technologies and provide an arc-extinguishing system, an independent three-phase switch, and a method for predicting the electrical lifetime of an independent three-phase switch based on a dual-arc-chamber synergistic degradation model. This method achieves a more accurate and reliable prediction of the remaining electrical lifetime of the three-phase switch by decomposing, monitoring, and modeling the key degradation parameters of each arc-extinguishing chamber and accurately quantifying their synergistic effects, thus providing a guarantee for the safe and stable operation of the power system.
[0006] To achieve the above objectives, the present invention provides the following technical solution:
[0007] In a first aspect, an arc-extinguishing system is provided, comprising a stationary contact, a first arc-extinguishing chamber, and a second arc-extinguishing chamber. The stationary contact has a stationary contact plate that cooperates with a moving contact and a terminal block. A mounting space is formed between the terminal block and the stationary contact plate. The second arc-extinguishing chamber is disposed within the mounting space. The second arc-extinguishing chamber includes a front magnetic plate and a rear magnetic plate. An arc-extinguishing cavity is formed between the front magnetic plate and the rear magnetic plate. An arc-extinguishing plate is disposed within the arc-extinguishing cavity. The arc-extinguishing plate includes a first arc-extinguishing plate located on the front magnetic plate and a second arc-extinguishing plate located on the rear magnetic plate. The arc extinguishing plates are arranged alternately, with the first and second arc extinguishing plates. A fixed frame is provided around the stationary contact plate to fix the first arc extinguishing chamber. The first arc extinguishing chamber includes several arc extinguishing grid plates and mounting plates arranged on both sides of the arc extinguishing grid plates. A gas generating hood is provided in the arc extinguishing space between the arc extinguishing grid plates. The gas generating hood is fixedly connected to the mounting plate on the same side. A placement cavity is formed inside the gas generating hood. A magnetic conductive sheet is provided in the placement cavity. The magnetic conductive sheet has a plug-in part in the direction of the fixed frame for connecting to the fixed frame.
[0008] In the above technical solution, a two-stage arc extinguishing mechanism is constructed by setting up a first arc extinguishing chamber and a second arc extinguishing chamber. When the moving and stationary contacts separate and generate an arc, the arc is first initially divided, cooled, and deionized between the staggered arc extinguishing plates in the second arc extinguishing chamber. Subsequently, the remaining arc energy is guided to the first arc extinguishing chamber, where it is further divided into a large number of short arcs in the arc extinguishing grid and finally extinguished. This staged treatment method of "preliminary treatment first, then complete extinguishing" avoids excessive pressure in a single arc extinguishing chamber, significantly improves the overall arc extinguishing capacity and breaking reliability, and is especially suitable for breaking large current and high energy arcs.
[0009] In another aspect, the present invention also provides an independent three-phase switch having the aforementioned arc-extinguishing system.
[0010] On the other hand, the present invention also provides a method for predicting the electrical lifetime of an independent three-phase switch, applicable to the aforementioned independent three-phase switch, comprising the following steps: constructing a comprehensive lifetime prediction model including a first arc-extinguishing chamber degradation sub-model, a second arc-extinguishing chamber degradation sub-model, and a synergistic effect model; performing multiple disconnection operations on a three-phase switch with dual arc-extinguishing chambers using an accelerated life test platform, the accelerated life test platform including an adjustable power supply, a triggering mechanism, a data acquisition system, and various online monitoring sensors; after each or every fixed number of disconnection operations, collecting a first set of characteristic parameters related to the first arc-extinguishing chamber, collecting a second set of characteristic parameters related to the second arc-extinguishing chamber, and collecting system-level synergistic parameters; inputting the first set, the second set of characteristic parameters, and the system-level synergistic parameters as inputs to the comprehensive lifetime prediction model; training the model using a machine learning algorithm to learn the degradation trajectory of each parameter and its interrelationships from the initial state to complete failure; finally, the model outputs the remaining electrical lifetime of the three-phase switch, expressed as the remaining number of reliable operations or a percentage of the health index.
[0011] The present invention further specifies that the establishment of the first arc-extinguishing chamber degradation sub-model is as follows: for the arc-extinguishing grid of the first arc-extinguishing chamber, an ablation model with accumulated arc energy as the independent variable is established, wherein the accumulated arc energy is obtained by integrating and summing the product of the arc current and arc voltage at each interruption; for the gas-generating shroud, a mass decay model with accumulated gas production as the independent variable is established, wherein the accumulated gas production is estimated by associating the peak gas pressure in the arc-extinguishing space with the interruption current level at each interruption; the outputs of the ablation model and the mass decay model are weighted and fused to generate a comprehensive index characterizing the health status of the first arc-extinguishing chamber, which monotonically decreases with the increase of the number of operations.
[0012] The present invention further specifies that the establishment of the second arc-extinguishing chamber degradation sub-model is as follows: for the staggered arc-extinguishing plates of the second arc-extinguishing chamber, the radius change and surface roughness evolution of their arc-shaped ends are quantified by periodically acquired three-dimensional scanning data, and a regression relationship with the number of breaks and the number of high-current breaks is established; for the carbon deposition inside the arc-extinguishing chamber, a carbon deposition index is established by analyzing the grayscale value and coverage area of the endoscopic images; the arc-extinguishing plate deformation is combined with the carbon deposition index, and a health status index characterizing the degradation of the insulation and magnetic blowout performance of the second arc-extinguishing chamber is generated by using principal component analysis to reduce the dimensionality.
[0013] The present invention further specifies that the synergistic effect model is used to quantify the mutual influence of the performance degradation of the first and second arc-extinguishing chambers: a quantitative relationship is established between the decrease in gas production efficiency of the first arc-extinguishing chamber and the decrease in airflow velocity within the arc-extinguishing chamber of the second arc-extinguishing chamber, characterized by calculating the deviation between the gas pressure rise rate and historical baseline data in each break test; a model is established to show that the deformation of the arc-extinguishing plate of the second arc-extinguishing chamber leads to changes in the magnetic circuit, which in turn affects the velocity at which the arc is introduced into the first arc-extinguishing chamber, quantified by analyzing the time change of the arc from the stationary contact to the grid plate of the first arc-extinguishing chamber recorded by high-speed imaging; the output of the synergistic effect model is a synergistic degradation factor, used to correct the independent prediction results of the degradation sub-models of the first and second arc-extinguishing chambers.
[0014] The invention further includes a mechanical life monitoring channel based on vibration signal analysis: an acceleration sensor is installed on the housing of the operating mechanism of the three-phase switch to collect vibration signals from each opening and closing operation; time-frequency analysis is performed on the vibration signals to extract the energy values of characteristic frequency components related to spring fatigue, lubrication deterioration, and component wear; a relationship model between the energy values of these characteristic frequency components and the number of operations is established and run in parallel with the electrical life prediction model; finally, a final remaining life prediction that comprehensively considers electrical life and mechanical life is given through decision-level fusion.
[0015] The present invention further includes a non-invasive online monitoring method for real-time updating of the prediction model: using a high-frequency current transformer to measure the high-frequency oscillation spectrum of the arc current during disconnection, it is found that as the arc-extinguishing grid ablation and the gas-generating shroud wear, the increase in arc instability leads to a significant enhancement of the spectral energy in a specific frequency band. The growth trend of this spectral energy can be used as an online proxy variable for the health status of the first arc-extinguishing chamber, and can be used to periodically update the parameters in the degradation sub-model without shutting down or disassembling the machine.
[0016] The present invention further includes an environment-based lifetime correction: temperature and humidity sensors are installed at key locations inside the three-phase switch to monitor environmental conditions in real time; an environmental correction factor function is established, which takes ambient temperature and relative humidity as input and outputs a lifetime decay acceleration coefficient; specifically, high temperature environments accelerate the aging of gas-generating materials and reduce insulation resistance, while high humidity environments increase the risk of electrochemical corrosion inside the arc-extinguishing chamber; the lifetime decay acceleration coefficient is multiplied by the lifetime prediction result based on the number of operations to obtain the remaining lifetime after environmental correction.
[0017] The present invention further includes a lifespan confidence interval assessment module: this module uses historical full-life test data of similar three-phase switches to generate a probability distribution of the actual remaining lifespan of the three-phase switch under different health status indices through Bootstrap or Monte Carlo simulation methods; for the currently assessed three-phase switch, based on its current health status index, the corresponding confidence interval is extracted from the probability distribution, and the prediction result is output in the form of "remaining number of operations: X times, confidence interval [Y, Z] times", thereby quantifying the uncertainty of the prediction.
[0018] The present invention further establishes a mapping relationship between the breaking current spectrum and the lifetime loss: the magnitude and timestamp of each current interrupted by the three-phase switch in actual operation are recorded; based on the different losses caused to the arc extinguishing system by different current levels, an equivalent lifetime loss unit is calculated for each breaking operation; the ELLU of all operations is accumulated to form the cumulative loss, and this is used as a more accurate input variable for the lifetime prediction model than based on the number of operations alone.
[0019] The present invention further specifies that the method is ultimately implemented through a digital twin: a one-to-one digital twin model is created for each physical three-phase switch on a cloud or edge computing node. This model integrates all the aforementioned degradation sub-models, synergy effect models, environmental correction models, and confidence interval assessment modules. The physical three-phase switch uploads key operational data to its digital twin through built-in sensors. The digital twin updates its own status in real time based on the received data and dynamically adjusts the lifespan prediction results. It then displays the current health status, predicted remaining lifespan, and maintenance recommendations to the user through a visual interface, thereby achieving predictive maintenance.
[0020] The above technical solution provides clear and reliable data support for maintenance decisions by quantifying the remaining number of reliable operations and confidence intervals. The method also has environmental adaptability. Through the equivalent lifetime loss unit (ELLU) model and environmental correction factor, the prediction results can dynamically adapt to different workloads and harsh environments, realizing intelligent management of the entire life cycle of the switch and ensuring the safety of the power system. Attached image description:
[0021] Figure 1 This is a schematic diagram of the structure of Embodiment 1 of the present invention.
[0022] Figure 2 This is a schematic diagram of the structure of the second arc-extinguishing chamber in Embodiment 1 of the present invention.
[0023] Figure 3 This is a schematic diagram of the structure of Embodiment 2 of the present invention.
[0024] Reference numerals: 10, stationary contact; 101, stationary contact plate; 102, terminal block; 20, first arc-extinguishing chamber; 30, second arc-extinguishing chamber; 40, front magnetic guide plate; 50, rear magnetic guide plate; 60, first arc-extinguishing plate; 70, second arc-extinguishing plate; 80, fixed frame; 90, arc-extinguishing grid plate; 100, mounting plate; 110, gas generating hood; 120, magnetic guide plate; 130, plug-in part; 140, slot. Detailed Implementation
[0025] The embodiments of this application will be described in detail below, providing a clear and complete description of the technical solutions within this application. Obviously, the described embodiments are merely a portion of the embodiments of this application, and not all of them. The components of this application described and shown herein can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.
[0026] Example 1
[0027] like Figure 1-2 As shown, this embodiment provides an arc-extinguishing system, including a stationary contact 10, a first arc-extinguishing chamber 20, and a second arc-extinguishing chamber 30. The stationary contact 10 has a stationary contact plate 101 that cooperates with a moving contact and a terminal block 102. A mounting space is formed between the terminal block 102 and the stationary contact plate 101. The second arc-extinguishing chamber 30 is disposed within the mounting space. The second arc-extinguishing chamber 30 includes a front magnetic guide plate 40 and a rear magnetic guide plate 50. An arc-extinguishing cavity is formed between the front magnetic guide plate 40 and the rear magnetic guide plate 50. An arc-extinguishing plate is disposed within the arc-extinguishing cavity. The arc-extinguishing plate includes a first arc-extinguishing plate 60 located on the front magnetic guide plate 40 and a second arc-extinguishing plate 70 located on the rear magnetic guide plate 50. The first arc-extinguishing plate 60 and the second arc-extinguishing plate 70 are connected. The arc plates 70 are staggered, and a fixed frame 80 is provided around the stationary contact plate 101. The fixed frame 80 is used to fix the first arc-extinguishing chamber 20. The first arc-extinguishing chamber 20 includes a plurality of arc-extinguishing grid plates 90 and mounting plates 100 provided on both sides of the arc-extinguishing grid plates 90. A gas generating hood 110 is provided in the arc-extinguishing space between the arc-extinguishing grid plates 90. The gas generating hood 110 is fixedly connected to the mounting plate 100 on the same side. A placement cavity is formed in the gas generating hood 110. A magnetic conductive sheet 120 is provided in the placement cavity. The magnetic conductive sheet 120 has a plug-in part 130 formed in the direction of the fixed frame 80 for connecting to the fixed frame 80. A slot 140 is formed on the fixed frame 80 to be inserted into the plug-in part 130.
[0028] This embodiment constructs a two-stage arc extinguishing mechanism by setting up a first arc-extinguishing chamber 20 and a second arc-extinguishing chamber 30. When the moving and stationary contacts 10 separate and generate an arc, the arc is first initially divided, cooled, and deionized between the staggered arc-extinguishing plates in the second arc-extinguishing chamber 30. Subsequently, the remaining arc energy is guided to the first arc-extinguishing chamber 20, where it is further divided into a large number of short arcs in the arc-extinguishing grid 90 and finally extinguished. This staged treatment method of "preliminary treatment followed by complete extinguishing" avoids excessive pressure in a single arc-extinguishing chamber, significantly improves the overall arc extinguishing capacity and breaking reliability, and is especially suitable for breaking large current and high-energy arcs.
[0029] Example 2
[0030] like Figure 3 As shown, this embodiment provides an independent three-phase switch with the arc extinguishing system described above.
[0031] Example 3
[0032] This embodiment provides a method for predicting the electrical lifetime of an independent three-phase switch, applicable to the aforementioned independent three-phase switch, comprising the following steps: constructing a comprehensive lifetime prediction model including a first arc-extinguishing chamber degradation sub-model, a second arc-extinguishing chamber degradation sub-model, and a synergistic effect model; performing multiple disconnection operations on a three-phase switch with dual arc-extinguishing chambers using an accelerated life testing platform, the accelerated life testing platform including an adjustable power supply, a triggering mechanism, a data acquisition system, and various online monitoring sensors; after each or every fixed number of disconnection operations, collecting a first set of characteristic parameters related to the first arc-extinguishing chamber, collecting a second set of characteristic parameters related to the second arc-extinguishing chamber, and collecting system-level synergistic parameters; inputting the first set, the second set of characteristic parameters, and the system-level synergistic parameters as inputs to the comprehensive lifetime prediction model; training the model using a machine learning algorithm to learn the degradation trajectory of each parameter and its interrelationships from the initial state to complete failure; finally, the model outputs the remaining electrical lifetime of the three-phase switch, expressed as the remaining number of reliable operations or a percentage of the health index.
[0033] The present invention further specifies that the establishment of the first arc-extinguishing chamber degradation sub-model is as follows: for the arc-extinguishing grid of the first arc-extinguishing chamber, an ablation model with accumulated arc energy as the independent variable is established, wherein the accumulated arc energy is obtained by integrating and summing the product of the arc current and arc voltage at each interruption; for the gas-generating shroud, a mass decay model with accumulated gas production as the independent variable is established, wherein the accumulated gas production is estimated by associating the peak gas pressure in the arc-extinguishing space with the interruption current level at each interruption; the outputs of the ablation model and the mass decay model are weighted and fused to generate a comprehensive index characterizing the health status of the first arc-extinguishing chamber, which monotonically decreases with the increase of the number of operations.
[0034] The present invention further specifies that the establishment of the second arc-extinguishing chamber degradation sub-model is as follows: for the staggered arc-extinguishing plates of the second arc-extinguishing chamber, the radius change and surface roughness evolution of their arc-shaped ends are quantified by periodically acquired three-dimensional scanning data, and a regression relationship with the number of breaks and the number of high-current breaks is established; for the carbon deposition inside the arc-extinguishing chamber, a carbon deposition index is established by analyzing the grayscale value and coverage area of the endoscopic images; the arc-extinguishing plate deformation is combined with the carbon deposition index, and a health status index characterizing the degradation of the insulation and magnetic blowout performance of the second arc-extinguishing chamber is generated by using principal component analysis to reduce the dimensionality.
[0035] The present invention further specifies that the synergistic effect model is used to quantify the mutual influence of the performance degradation of the first and second arc-extinguishing chambers: a quantitative relationship is established between the decrease in gas production efficiency of the first arc-extinguishing chamber and the decrease in airflow velocity within the arc-extinguishing chamber of the second arc-extinguishing chamber, characterized by calculating the deviation between the gas pressure rise rate and historical baseline data in each break test; a model is established to show that the deformation of the arc-extinguishing plate of the second arc-extinguishing chamber leads to changes in the magnetic circuit, which in turn affects the velocity at which the arc is introduced into the first arc-extinguishing chamber, quantified by analyzing the time change of the arc from the stationary contact to the grid plate of the first arc-extinguishing chamber recorded by high-speed imaging; the output of the synergistic effect model is a synergistic degradation factor, used to correct the independent prediction results of the degradation sub-models of the first and second arc-extinguishing chambers.
[0036] The invention further includes a mechanical life monitoring channel based on vibration signal analysis: an acceleration sensor is installed on the housing of the operating mechanism of the three-phase switch to collect vibration signals from each opening and closing operation; time-frequency analysis is performed on the vibration signals to extract the energy values of characteristic frequency components related to spring fatigue, lubrication deterioration, and component wear; a relationship model between the energy values of these characteristic frequency components and the number of operations is established and run in parallel with the electrical life prediction model; finally, a final remaining life prediction that comprehensively considers electrical life and mechanical life is given through decision-level fusion.
[0037] The present invention further includes a non-invasive online monitoring method for real-time updating of the prediction model: using a high-frequency current transformer to measure the high-frequency oscillation spectrum of the arc current during disconnection, it is found that as the arc-extinguishing grid ablation and the gas-generating shroud wear, the increase in arc instability leads to a significant enhancement of the spectral energy in a specific frequency band. The growth trend of this spectral energy can be used as an online proxy variable for the health status of the first arc-extinguishing chamber, and can be used to periodically update the parameters in the degradation sub-model without shutting down or disassembling the machine.
[0038] The present invention further includes an environment-based lifetime correction: temperature and humidity sensors are installed at key locations inside the three-phase switch to monitor environmental conditions in real time; an environmental correction factor function is established, which takes ambient temperature and relative humidity as input and outputs a lifetime decay acceleration coefficient; specifically, high temperature environments accelerate the aging of gas-generating materials and reduce insulation resistance, while high humidity environments increase the risk of electrochemical corrosion inside the arc-extinguishing chamber; the lifetime decay acceleration coefficient is multiplied by the lifetime prediction result based on the number of operations to obtain the remaining lifetime after environmental correction.
[0039] The present invention further includes a lifespan confidence interval assessment module: this module uses historical full-life test data of similar three-phase switches to generate a probability distribution of the actual remaining lifespan of the three-phase switch under different health status indices through Bootstrap or Monte Carlo simulation methods; for the currently assessed three-phase switch, based on its current health status index, the corresponding confidence interval is extracted from the probability distribution, and the prediction result is output in the form of "remaining number of operations: X times, confidence interval [Y, Z] times", thereby quantifying the uncertainty of the prediction.
[0040] The present invention further establishes a mapping relationship between the breaking current spectrum and the lifetime loss: the magnitude and timestamp of each current interrupted by the three-phase switch in actual operation are recorded; based on the different losses caused to the arc extinguishing system by different current levels, an equivalent lifetime loss unit is calculated for each breaking operation; the ELLU of all operations is accumulated to form the cumulative loss, and this is used as a more accurate input variable for the lifetime prediction model than based on the number of operations alone.
[0041] The present invention further specifies that the method is ultimately implemented through a digital twin: a one-to-one digital twin model is created for each physical three-phase switch on a cloud or edge computing node. This model integrates all the aforementioned degradation sub-models, synergy effect models, environmental correction models, and confidence interval assessment modules. The physical three-phase switch uploads key operational data to its digital twin through built-in sensors. The digital twin updates its own status in real time based on the received data and dynamically adjusts the lifespan prediction results. It then displays the current health status, predicted remaining lifespan, and maintenance recommendations to the user through a visual interface, thereby achieving predictive maintenance.
[0042] The above technical solution provides clear and reliable data support for maintenance decisions by quantifying the remaining number of reliable operations and confidence intervals. The method also has environmental adaptability. Through the equivalent lifetime loss unit (ELLU) model and environmental correction factor, the prediction results can dynamically adapt to different workloads and harsh environments, realizing intelligent management of the entire life cycle of the switch and ensuring the safety of the power system.
[0043] The following is a further explanation of this embodiment:
[0044] First, the accelerated life testing platform was built: this platform is the foundation for acquiring model training data and conducting validation. It includes:
[0045] Adjustable three-phase power supply: simulates various fault current conditions in the actual power grid, with the current range covering from normal operating current to ultimate breaking current.
[0046] Precision triggering mechanism: controls the moving contacts of the three-phase switch to perform precise and repeatable opening and closing operations, and records mechanical parameters such as breaking speed and contact force.
[0047] Integrated data acquisition system: The core component includes a high-speed data acquisition card (sampling rate not less than 10MS / s), a synchronous trigger, and various sensors.
[0048] Sensor array:
[0049] Electrical sensors: High-voltage differential voltage probes and Rogowski coil current sensors are used to accurately measure the arc voltage and arc current waveforms when each phase is interrupted.
[0050] Optical sensor: High-speed camera with a frame rate of no less than 10,000 fps, used to observe the trajectory and speed of the electric arc moving from the stationary contact to the first arc-extinguishing chamber.
[0051] Pressure sensor: A high-frequency response miniature pressure sensor, installed on a fixed frame or in the airflow channel between arc-extinguishing chambers, is used to capture transient pressure changes in the airflow during the interruption process.
[0052] Vibration sensor: accelerometer, installed on the housing of the operating mechanism, to collect vibration signals during the opening and closing of the circuit breaker.
[0053] Environmental sensor: Temperature and humidity sensor, installed inside the switch housing, to monitor ambient temperature and humidity.
[0054] High-frequency current transformer: used to detect the 1-10MHz high-frequency oscillation component in arc current.
[0055] Data collection process:
[0056] Periodic offline precision measurement: In order to obtain direct degradation data of the internal components of the arc extinguishing chamber, it is necessary to periodically (e.g., after every 100 break tests) shut down and partially disassemble the switch.
[0057] The mass of the gas-generating shroud in the first arc-extinguishing chamber was weighed using a precision balance, and its mass loss rate (ΔM) was calculated.
[0058] A 3D scanner was used to scan the staggered arc-extinguishing discs in the second arc-extinguishing chamber, quantifying the radius variation (ΔR) and surface roughness (Ra) of their arc-shaped ends. An endoscope and an industrial camera were used to photograph the interior of the arc-extinguishing chamber of the second arc-extinguishing chamber, and the carbon deposition coverage (C_cov) was calculated using image processing algorithms (such as grayscale statistics and edge recognition).
[0059] Continuous online monitoring: During each interruption operation, all online sensors are simultaneously triggered to collect data.
[0060] Record the arc voltage V(t) and arc current I(t), and calculate the arc energy at each break. And arc ignition time.
[0061] Record the airflow pressure waveform P(t) and calculate its rate of rise (dp / dt).
[0062] Record vibration signals and high-frequency current spectra.
[0063] Record the ambient temperature (T) and relative humidity (RH).
[0064] Feature extraction and health index construction:
[0065] The health index (H1) of the first arc-extinguishing chamber was constructed as follows:
[0066] Inputs: Arc extinguishing grid ablation data (indirectly estimated through arc energy accumulation or offline measurement), gas generation shroud mass loss (ΔM), and online arc energy (E_arc).
[0067] deal with:
[0068] An arc-extinguishing grid ablation model was established: the reduction in grid thickness (ΔT) is positively correlated with the cumulative arc energy (ΣE_arc), which was obtained by fitting historical data.
[0069] A mass decay model for the gas generation hood was established: the mass loss (ΔM) is positively correlated with the estimated cumulative gas generation (derived from the correlation between peak gas pressure and breaking current level).
[0070] Fusion: A weighted fusion algorithm is used, for example: in, and As the initial value, Let H1 be the baseline energy, and α, β, γ be weighting coefficients determined through expert experience or regression analysis, with α + β + γ = 1. H1 is an exponential function that monotonically decreases from 1 (healthy) to 0 (failure).
[0071] Construction of the health index (H2) of the second arc-extinguishing chamber:
[0072] Inputs: Arc extinguishing plate shape variables (ΔR, Ra), carbon deposition index (C_cov).
[0073] Processing: Since ΔR, Ra and C_cov may have multicollinearity, principal component analysis (PCA) is used to reduce the dimensionality and decorrelate these parameters to extract 1-2 principal components.
[0074] The principal component scores are mapped to the [0,1] interval using a function (such as the Sigmoid function or a linear combination) to generate the H2 index. A decrease in H2 indicates a degradation in the insulation performance and magnetic blowout arc guiding capability of the second arc-extinguishing chamber. The method for obtaining the H2 index is as follows: Z-score standardization is performed on ΔR, Ra, and C_cov. The standardized parameters are then combined into an n×3 matrix X, where n is the number of samples. The covariance matrix Σ is calculated, and the characteristic equation |Σ - λI| = 0 is solved to obtain the eigenvalues λ1, λ2, λ3 (arranged in descending order) and the corresponding eigenvectors v1, v1, v3. The standardized data is then projected onto the selected principal components, with the selection based on the cumulative contribution rate. Then calculate the principal component scores. The following result is generated through linear mapping:
[0075] ; and These are the minimum and maximum baseline values of the principal component scores, respectively;
[0076] Alternatively, the following result can be generated through Sigmoid mapping:
[0077] .
[0078] Calculation of the co-degradation factor (λ):
[0079] Inputs: historical baseline and current values of airflow pressure rise rate (dp / dt), and arc motion time measured by high-speed camera.
[0080] Process: Calculate the pressure rise rate ratio , The decrease directly reflects the negative impact of the reduced gas production capacity of the first arc-extinguishing chamber on the airflow environment of the second arc-extinguishing chamber; among them, This is the rate of increase in gas pressure inside the arc-extinguishing chamber, measured during a breaking test when the circuit breaker is in brand-new condition. This is a baseline value. The rate of increase in air pressure is measured when the same test is performed with the circuit breaker in its current state.
[0081] Calculate the arc motion time ratio . The increase reflects the change in magnetic circuit caused by the deformation of the arc-extinguishing plate in the second arc-extinguishing chamber, which weakens the ability to push the arc towards the first arc-extinguishing chamber. In a brand new state, the time elapsed from the generation of an electric arc between the contacts to its complete absorption into the grid of the first arc-extinguishing chamber. : The time taken for the same process under the current conditions.
[0082] The co-degradation factor λ is defined as: ,in and These are the weighting coefficients. δ + θ =1, This indicates good synergy; This indicates a decline in synergistic effect.
[0083] Model Training and Lifetime Prediction: Comprehensive Model Training:
[0084] Model selection: A machine learning model suitable for processing time series data, such as a Long Short-Term Memory (LSTM) network, is adopted. Its input is a time series window, and the data in the window includes [H1, H2, λ, vibration characteristics, ambient temperature, ambient humidity, equivalent lifetime loss units (ELLU)].
[0085] Training data: By conducting accelerated full-lifecycle tests on multiple three-phase switches of the same model from brand new to failure, all characteristic data and the remaining lifespan percentage (label) corresponding to the timestamps throughout the entire lifecycle are collected.
[0086] Training process: The LSTM network is supervised learning using a large-scale dataset to enable the model to learn to map input feature sequences to the remaining lifespan percentage.
[0087] Online forecasts and updates:
[0088] For the online running switch to be predicted, the latest extracted feature sequence is periodically (e.g., every 50 operations) fed into the trained LSTM model.
[0089] The model outputs its remaining number of reliable operations and percentage of health status.
[0090] Meanwhile, the confidence interval assessment module will call historical failure distribution data and give a prediction interval (for example, "1500 remaining failures, 90% confidence interval is [1300, 1700] failures") to quantify the uncertainty of the prediction.
[0091] Non-intrusive update: Using high-frequency spectrum energy as a proxy variable for H1, the model input is fine-tuned without disassembling the model, thus achieving online update of the model state.
[0092] Multi-information fusion:
[0093] Mechanical life channel: The mechanical life prediction provided by vibration signal analysis is fused with the electrical life prediction results at the decision level through DS evidence theory to give a comprehensive remaining life.
[0094] Environmental correction: The real-time monitored temperature and humidity are substituted into a predefined acceleration factor function to correct the predicted number of operations, making it more consistent with the actual operating environment.
[0095] Load spectrum mapping: By using the ELLU model, different breaking currents are converted into a unified loss unit, making the prediction model closer to the actual working load of the switch.
[0096] Digital Twin Implementation: The final form of the method of this invention is implemented through a digital twin. A virtual image is created for each physical three-phase switch on a cloud platform or edge server. This digital twin integrates all the aforementioned models, algorithms, and data processing logic. The physical switch, through its built-in IoT module, encrypts and uploads key operational data (breaking current, success flags, environmental data, etc.) to its twin. The twin performs real-time calculations, dynamically updates its lifespan prediction, and intuitively displays the switch's real-time health status, degradation trend, predicted remaining lifespan, and specific maintenance recommendations (such as "The first arc-extinguishing chamber gas generation hood is severely worn; planned maintenance is recommended") to the user via a web interface or mobile app. Ultimately, this achieves an intelligent upgrade from "post-fault maintenance" to "preventive maintenance" and then to "predictive maintenance."
[0097] In addition, the Long Short-Term Memory (LSTM) network model is used for comprehensive lifetime prediction, and the specific method is as follows: It includes the following network structure:
[0098] Input layer: The input is a time series window Each of them It is a feature vector containing H1, H2, λ, vibrational characteristic energy, ambient temperature, ambient humidity, ELLU, etc., collected at time point i. The window size n is a hyperparameter, usually chosen based on the data sampling frequency and the inertia of the degradation process, for example, n=20 (representing data from the most recent 20 operations).
[0099] Hidden layers: One or more LSTM units are used. Each LSTM unit contains an input gate, a forget gate, an output gate, and a cell state, which can effectively learn long-term dependencies in time series. The number of hidden layer nodes (e.g., 64, 128) needs to be adjusted experimentally.
[0100] Output layer: Typically a fully connected layer followed by an activation function. For regression prediction (remaining lifetime percentage), a linear activation function or a sigmoid function (constraining the output between 0 and 1) can be used. If it is necessary to predict the remaining number of operations, the output can be inversely normalized.
[0101] Loss Function: The mean squared error (MSE) or mean absolute error (MAE) is used as the loss function to measure the difference between the model's predicted percentage of remaining lifespan and the true value. Optimizer: The Adam (Adaptive Moment Estimation) optimizer is used, which adaptively adjusts the learning rate, resulting in high training efficiency. Training Process: 70% of the full-lifespan trial dataset is used as the training set, 15% as the validation set to tune hyperparameters and prevent overfitting, and the final 15% as the test set to evaluate the final performance. Training will be performed in multiple epochs until the model's loss on the validation set no longer decreases significantly.
[0102] DS evidence theory applied to electromechanical life fusion:
[0103] Recognition Framework Establishment: Define the recognition framework Θ = {L, H}, where L represents "lifespan nearing its end" and H represents "healthy lifespan". Its power set is {∅, {L}, {H}, {L, H}}. {L, H} represents "uncertainty".
[0104] Basic Probability Assignment (BPA):
[0105] Electrical lifetime channel: Assign BPA to the remaining lifetime percentage P_e predicted by the LSTM model. For example: m_e({L}) = 1-P_e, m_e({H}) = P_e * 0.8, m_e({L, H}) = P_e * 0.2 (indicating a 20% uncertainty regarding the health status).
[0106] Mechanical life channel: Based on the relationship model between vibration characteristics and number of operations, a mechanical health index P_m is derived, and BPA is similarly assigned: m_m({L}) = 1 - P_m, m_m({H}) = P_m * 0.7, m_m({L, H}) = P_m * 0.3.
[0107] Dempster's combination rule: Combine the BPA (Body Aspects) of two sources of evidence (electric and mechanical). The combination rule is as follows:
[0108] ; ;
[0109] , is a normalization constant used to exclude conflicting evidence.
[0110] For example, calculate the confidence of "life expectancy and health" {H} after fusion: m_fused({H}) = [m_e({H}) * m_m({H}) + m_e({H}) * m_m({L,H}) + m_e({L,H}) * m_m({H})] / K;
[0111] Decision: Based on the fused BPA, select the proposition with the highest reliability as the final judgment. For example, if m_fused({H}) > m_fused({L}), then the overall switch is judged to be healthy.
[0112] Equivalent Lifetime Loss Unit (ELLU) Model: Principle: Based on the ablation effect of arc energy on the arc-extinguishing chamber material, it approximately follows the principle of I... 2 t-law.
[0113] ; Let be the effective value current of the i-th interruption operation. The arcing time of the i-th interruption operation; The selected reference current (such as the rated operating current). The baseline arcing time;
[0114] Cumulative Losses: The total cumulative losses of the switch are the sum of the ELLU values of all historical disconnect operations.
[0115] As a key input feature of the LSTM model, it reflects the actual electrical stress experienced by the switch more accurately than the number of operations N alone.
[0116] Bootstrap-based confidence interval assessment
[0117] Bootstrap resampling: From the historical full-lifetime test dataset (containing the mapping relationships of all H1, H2 -> actual lifespan of multiple switches from new to bad), a subset of samples of the same size as the original dataset is randomly drawn with replacement. This process is repeated B times (e.g., B=1000) to obtain B Bootstrap sample sets.
[0118] Model Training and Prediction: On each Bootstrap sample set, retrain the lifespan prediction model (such as a simplified regression model). For the switch to be evaluated, input its current H1 and H2 into these B models to obtain B remaining lifespan predictions.
[0119] Confidence interval generation: Sort the B predicted values from smallest to largest. Then take the 5th percentile and the 95th percentile, which constitute the 90% confidence interval for the remaining lifetime of the switch in its current state. For example, after sorting 1000 predicted values, the 50th value is the lower limit of the interval, and the 950th value is the upper limit. This means we have 90% confidence that the actual remaining lifetime of the switch will fall within this interval.
[0120] Environmental correction factor function
[0121] The Arrennis model (for temperature): used to simulate the accelerating effect of high temperatures on the aging and insulation degradation of gas-generating materials. , As an acceleration factor; The activation energy of the material is given by ; k is the Boltzmann constant. The absolute temperature of the environment in which the arc-extinguishing chamber actually operates; The reference temperature for lifetime prediction and comparison.
[0122] Parker model (for humidity): used to simulate the accelerating effect of high humidity on electrochemical corrosion. , As a humidity accelerating factor, This represents the actual operating humidity. For reference humidity; n=3.
[0123] Comprehensive correction: Total environmental acceleration factor The corrected remaining lifetime is: (model-predicted remaining lifetime) / .
[0124] If certain terms are used in the specification and claims to refer to specific components, those skilled in the art will understand that hardware manufacturers may use different names to refer to the same component. This specification and claims do not distinguish components based on differences in name, but rather on differences in function. The term "comprising" as used throughout the specification and claims is an open-ended term and should be interpreted as "comprising but not limited to." "Approximately" means that within an acceptable margin of error, those skilled in the art can solve the technical problem and substantially achieve the technical effect within a certain margin of error.
[0125] It should be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a product or system comprising a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a product or system. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the product or system that includes that element.
[0126] The foregoing description illustrates and describes several preferred embodiments of the present invention. However, as previously stated, it should be understood that the present invention is not limited to the forms disclosed herein and should not be construed as excluding other embodiments. It can be used in various other combinations, modifications, and environments, and can be altered within the scope of the inventive concept by means of the foregoing teachings or techniques or knowledge in related fields. Any modifications and variations made by those skilled in the art that do not depart from the spirit and scope of the present invention should be within the protection scope of the appended claims.
Claims
1. A method for predicting the electrical life of an independent three-phase switch, wherein the independent three-phase switch includes an arc-extinguishing system, the arc-extinguishing system including a stationary contact, a first arc-extinguishing chamber, and a second arc-extinguishing chamber, the stationary contact having a stationary contact plate that mates with a moving contact and a terminal block, a mounting space being formed between the terminal block and the stationary contact plate, the second arc-extinguishing chamber being disposed within the mounting space, the second arc-extinguishing chamber including a front magnetic plate and a rear magnetic plate, an arc-extinguishing cavity being formed between the front magnetic plate and the rear magnetic plate, an arc-extinguishing plate being disposed within the arc-extinguishing cavity, the arc-extinguishing plate including a first arc-extinguishing plate located on the front magnetic plate and a second arc-extinguishing plate located on the front magnetic plate. A second arc-extinguishing plate is located on the rear magnetic guide plate. The first and second arc-extinguishing plates are staggered. A fixed frame is provided around the stationary contact plate to fix the first arc-extinguishing chamber. The first arc-extinguishing chamber includes a plurality of arc-extinguishing grid plates and mounting plates disposed on both sides of the arc-extinguishing grid plates. A gas-generating hood is disposed in the arc-extinguishing space between the arc-extinguishing grid plates. The gas-generating hood is fixedly connected to the mounting plate on the same side. A placement cavity is formed inside the gas-generating hood. A magnetic guide plate is disposed in the placement cavity. The magnetic guide plate has an insertion part forming towards the fixed frame for connecting to the fixed frame. The method includes the following steps: constructing a comprehensive lifetime prediction model comprising a first arc-extinguishing chamber degradation sub-model, a second arc-extinguishing chamber degradation sub-model, and a synergistic effect model; performing multiple disconnection operations on a three-phase switch with dual arc-extinguishing chambers using an accelerated life testing platform, the accelerated life testing platform including an adjustable power supply, a triggering mechanism, a data acquisition system, and various online monitoring sensors; after each or every fixed number of disconnection operations, collecting a first set of characteristic parameters related to the first arc-extinguishing chamber, collecting a second set of characteristic parameters related to the second arc-extinguishing chamber, and collecting system-level synergistic parameters; inputting the first set, the second set of characteristic parameters, and the system-level synergistic parameters into the comprehensive lifetime prediction model; training the comprehensive lifetime prediction model using a machine learning algorithm to learn the degradation trajectory of each parameter and its interrelationships from the initial state to complete failure; finally, the comprehensive lifetime prediction model outputs the remaining electrical lifetime of the three-phase switch, expressed as the remaining number of reliable operations or a percentage of the health index.
2. The method according to claim 1, characterized in that, The establishment of the first arc-extinguishing chamber degradation sub-model is as follows: For the arc-extinguishing grid of the first arc-extinguishing chamber, an ablation model with accumulated arc energy as the independent variable is established. The accumulated arc energy is obtained by integrating and summing the product of the arc current and arc voltage at each interruption. For the gas-generating shroud, a mass decay model with accumulated gas generation as the independent variable is established. The accumulated gas generation is estimated by associating the peak gas pressure in the arc-extinguishing space with the interruption current level at each interruption. The outputs of the ablation model and the mass decay model are weighted and fused to generate a comprehensive index characterizing the health status of the first arc-extinguishing chamber. This index decreases monotonically with the increase of the number of operations.
3. The method according to claim 1, characterized in that, The establishment of the second arc-extinguishing chamber degradation sub-model is as follows: For the staggered arc-extinguishing plates of the second arc-extinguishing chamber, the radius change and surface roughness evolution of their arc-shaped ends are quantified by periodically acquired three-dimensional scanning data, and a regression relationship with the number of breaks and the number of high-current breaks is established; For the carbon deposition inside the arc-extinguishing chamber, a carbon deposition index is established by analyzing the gray value and coverage area of the endoscopic images; The arc-extinguishing plate shape variable is combined with the carbon deposition index, and a health status index characterizing the degradation of the insulation and magnetic blowout performance of the second arc-extinguishing chamber is generated by using principal component analysis to reduce the dimensionality.
4. The method according to claim 1, characterized in that, The synergistic effect model is used to quantify the mutual influence of the performance degradation of the first and second arc-extinguishing chambers: a quantitative relationship is established between the decrease in gas production efficiency of the first arc-extinguishing chamber and the decrease in airflow velocity in the arc-extinguishing chamber of the second arc-extinguishing chamber, which is characterized by calculating the deviation of the gas pressure rise rate from historical baseline data in each break test; a model is established to show that the deformation of the arc-extinguishing plate of the second arc-extinguishing chamber leads to changes in the magnetic circuit, which in turn affects the speed at which the arc is introduced into the first arc-extinguishing chamber, which is quantified by analyzing the time change of the arc from the stationary contact to the grid plate of the first arc-extinguishing chamber recorded by high-speed camera; the output of the synergistic effect model is a synergistic degradation factor, which is used to correct the independent prediction results of the degradation sub-models of the first and second arc-extinguishing chambers.
5. The method according to claim 1, characterized in that, It also includes a mechanical life monitoring channel based on vibration signal analysis: an acceleration sensor is installed on the housing of the operating mechanism of the three-phase switch to collect vibration signals for each opening and closing operation; time-frequency analysis is performed on the vibration signals to extract the energy values of characteristic frequency components related to spring fatigue, lubrication deterioration, and component wear; a relationship model between the energy values of these characteristic frequency components and the number of operations is established and run in parallel with the electrical life prediction model; finally, a final remaining life prediction that comprehensively considers electrical life and mechanical life is given through decision-level fusion.
6. The method according to claim 1, characterized in that, The method introduces a non-invasive online monitoring approach for real-time updating of the prediction model: using a high-frequency current transformer to measure the high-frequency oscillation spectrum of the arc current during disconnection, it is found that as the arc-extinguishing grid ablation and the gas-generating shroud wear, the increase in arc instability leads to a significant enhancement of the spectral energy in a specific frequency band. The growth trend of this spectral energy serves as an online proxy variable for the health status of the first arc-extinguishing chamber, used to periodically update the parameters in the degradation sub-model without shutting down the machine or disassembling it.
7. The method according to claim 1, characterized in that, Environmental-based lifetime correction is performed: temperature and humidity sensors are installed inside the three-phase switch to monitor environmental conditions in real time; an environmental correction factor function is established, which takes ambient temperature and relative humidity as input and outputs a lifetime decay acceleration coefficient; specifically, high temperature environment will accelerate the aging of gas-generating materials and reduce insulation resistance, and high humidity environment will increase the risk of electrochemical corrosion inside the arc-extinguishing chamber; the lifetime decay acceleration coefficient is multiplied by the lifetime prediction result based on the number of operations to obtain the remaining lifetime after environmental correction.
8. The method according to claim 7, characterized in that, The method includes a lifetime confidence interval assessment module: This module uses historical full-life test data of similar three-phase switches to generate a probability distribution of the actual remaining lifetime of the three-phase switch under different health status indices through Bootstrap or Monte Carlo simulation methods; For the currently evaluated three-phase switch, based on its current health status index, the corresponding confidence interval is extracted from the probability distribution, and the prediction result is output in the form of "remaining number of operations: X times, confidence interval [Y, Z] times", thereby quantifying the uncertainty of the prediction.
9. The method according to claim 8, characterized in that, Establish a mapping relationship between breaking current spectrum and lifetime loss: record the magnitude and timestamp of each current interrupted by the three-phase switch in actual operation; calculate an equivalent lifetime loss unit for each breaking operation based on the different degree of loss caused to the arc extinguishing system by different current levels; accumulate the equivalent lifetime loss units of all operations to form cumulative loss, and use it as a more accurate input variable for lifetime prediction model than based on the number of operations alone.
10. The method according to claim 9, characterized in that, The method is ultimately implemented through a digital twin: a one-to-one digital twin model is created for each physical three-phase switch on the cloud or edge computing node. This model integrates all the aforementioned degradation sub-models, synergy effect models, environmental correction models, and confidence interval assessment modules. The physical three-phase switch uploads operational data to its digital twin through built-in sensors. The digital twin updates its own status in real time based on the received data and dynamically adjusts the lifespan prediction results. It then displays the current health status, predicted remaining lifespan, and maintenance recommendations to the user through a visual interface, thus achieving predictive maintenance.