Multi-source signal fused intelligent monitoring and fault adaptive protection system for high-voltage switch cabinet

Abstract

The invention discloses a multi-source signal fusion high-voltage switch cabinet intelligent monitoring and fault self-adaptive protection system, and relates to the technical field of switch cabinets, the system comprises a multi-source sensing module, a state evaluation module and a self-adaptive protection module, the multi-source sensing module synchronously collects UHF, AE, arc light and primary loop current, completes front-end conditioning and unifies time reference, and the state evaluation module evaluates the state of the state evaluation module and the self-adaptive protection module; and as a basis for subsequent association and calculation, the state evaluation module completes grading based on a double threshold + trend, establishes and calls a nonlinear relationship between partial discharge and overcurrent according to a baseline of the same model, and calculates residual error and durability, an output grade, a high-risk sign and an auxiliary criterion on line. The self-adaptive protection module executes arc light correlation partial discharge and arc light overcurrent fusion triggering, tripping is carried out when the fusion criterion is met, sensitivity enhancement is carried out on a correlation threshold and a front window when the residual error continuously exceeds the threshold, and the method has the advantage of being accurate in judgment.

Description

Technical Field

[0001] This invention relates to the field of switchgear technology, specifically to a high-voltage switchgear intelligent monitoring and fault adaptive protection system that integrates multi-source signals. Background Technology

[0002] High-voltage switchgear is the core equipment of the power distribution network, and its internal faults, especially insulation faults, are the main cause of serious accidents. In-depth analysis shows that arcing faults inside the cabinet mainly stem from two types of causes: first, insulation breakdown caused by long-term deterioration of insulation materials; second, transient short circuits caused by non-insulation material problems such as loose mechanical connections and foreign object intrusion.

[0003] Existing technologies include partial discharge online monitoring systems that focus on monitoring early signs of Type I faults (insulation degradation) using sensors such as UHF and ultrasonic (AE) sensors to achieve predictive maintenance; and arc flash protection devices that commonly use a dual criterion of arc flash and overcurrent to quickly cut off power after an arc fault occurs, thus providing post-fault protection. However, the former cannot achieve rapid tripping protection at the moment of the fault and is powerless against Type II sudden faults. Sometimes, it may misjudge similar characteristics of corona discharge, external electromagnetic radiation, mechanical and airflow noise as partial discharge, resulting in false tripping. The latter cannot distinguish the cause of arc flash. For Type I faults, it is susceptible to false tripping due to light interference due to the lack of partial discharge criteria, and for high-impedance Type II faults, it may fail to trip due to insufficient current. Therefore, it is necessary to design a multi-source signal fusion intelligent monitoring and fault adaptive protection system for high-voltage switchgear. Summary of the Invention

[0004] The purpose of this invention is to provide a high-voltage switchgear intelligent monitoring and fault adaptive protection system based on multi-source signal fusion, so as to solve the problems mentioned in the background art.

[0005] To address the aforementioned technical problems, this invention provides the following technical solution: a multi-source signal fusion intelligent monitoring and fault adaptive protection system for high-voltage switchgear, comprising a multi-source sensing module, a state assessment module, and an adaptive protection module. The multi-source sensing module synchronously collects UHF, AE, arc, and primary circuit current, completing front-end conditioning and unifying the time reference as the basis for subsequent correlation and calculation. The state assessment module completes classification based on dual thresholds and trends, and establishes and calls the nonlinear relationship between partial discharge and overcurrent according to the same model baseline, calculates residuals and persistence online, and outputs the level, high-risk indicator, and auxiliary criteria. The adaptive protection module performs arc-related partial discharge and arc-related overcurrent fusion triggering, tripping when the fusion criteria are met, and enhancing the sensitivity of the correlation threshold and pre-window when the residual continuously exceeds the threshold.

[0006] According to the above technical solution, the multi-source sensing module includes a multi-source sensing unit and a fusion judgment module. The multi-source sensing unit consists of a UHF sensor, an ultrasonic sensor, an arc light sensor, and a current sampling module, used to collect four signals: UHF, AE, arc light, and current. The fusion judgment module performs PRPD spectrum analysis on the partial discharge signal to extract the characteristics of discharge quantity and discharge frequency; and calculates the effective value of the current signal. The state assessment module includes a dual-threshold trend enhancement grading module, a nonlinear baseline modeling module, and a residual determination module. The dual-threshold trend enhancement grading module is used to output normal, first-level, and second-level grades with low and high threshold superimposed trends and persistence rules, focusing on sensitive identification of continuous degradation. The nonlinear baseline modeling module establishes and calls nonlinear mappings of partial discharge and overcurrent for each model, providing expected changes in load interpretability. The residual determination module calculates whether the residual determination of actual and expected changes continues to deviate. The adaptive protection module includes an association threshold adjustment module, a trip control module, and an event packaging and reporting module. The association threshold adjustment module lowers the partial discharge association threshold by a coefficient when nonlinear inconsistency occurs. The event packaging and reporting module sends various data as structured evidence along with the level and action results. The trip control module is used to trip when a fault occurs.

[0007] Based on the above technical solution, the working method of this system is as follows: S1. Simultaneously acquire four signals: UHF, AE, arc, and current, and complete front-end analog and digital conditioning. Generate pulse sequences and amplitude statistics for the partial discharge channel that can be used for PRPD analysis. Calculate the effective value and over-limit flag for the current. Align the timestamps of each channel and establish a consistent time reference for subsequent correlation of partial discharge, arc, and overcurrent. The current is the operating current of the primary circuit. S2. Based on the partial discharge characteristic value and trend, the system is classified into normal, first-level early warning and second-level alarm levels, respectively, and corresponding prompts for attention, suggestions for enhanced maintenance and local audible and visual reporting are provided. The threshold adopts a dual-threshold structure and combines trend to enhance the sensitivity to continuous deterioration. When the second-level alarm state is entered, the system marks the insulation as high risk, providing a priori conditions for the sensitive switching of the protection logic. S3. The triggering criteria adopt configurable logic of arc-related partial discharge and arc-related overcurrent to achieve both prevention of false tripping and prevention of failure to trip. The insulation deterioration type fault is triggered by the arc-related partial discharge signal to suppress false tripping caused by pure optical interference. The non-insulation deterioration type sudden short circuit is backed by arc-related + overcurrent to ensure that the fault is cleared without missing detection. When the fusion criteria are met, a trip command is immediately output to clear the fault. S4. For the same type of switchgear, based on the collected partial discharge signal and current RMS value, establish a nonlinear correspondence between the time change rate of the partial discharge comprehensive index and the time change rate of the current RMS value, and form a baseline for this type of switchgear. S5. During operation, both are calculated synchronously at the same time step and substituted into the baseline relationship to obtain the expected change. If the deviation between the actual change and the expected value continues to exceed the threshold, it is judged as an abnormal sign that is inconsistent with simple load fluctuations, and is used as an auxiliary criterion for alarm and protection triggering.

[0008] According to the above technical solution, the alarm classification in S2 is specifically as follows: The device provides amplitude statistics for the partial discharge channels UHF and AE. With pulse statistics Threshold discrimination is performed in a sliding window. The amplitudes of AE and UHF are taken respectively. With pulse counting ,calculate , ,in , These are the normal amplitude threshold and the high amplitude threshold, respectively. , These represent the normal and high thresholds for pulse counting, and the comprehensive partial discharge index. , For the linear contribution of the pulse, The amplification factor is used to classify the system into three levels: normal, first-level warning, and second-level alarm. A value below the preset alarm threshold is marked as normal; a value exceeding the lower threshold is marked as... However, a Level 1 warning is issued when the development trend is stable; otherwise, a high threshold is exceeded. And when it shows a continuous increasing trend, that is The system triggers a level 2 alarm when the value increases within multiple consecutive sliding windows.

[0009] According to the above technical solution, the triggering criterion in S3 is specifically as follows: S3-1. For insulation degradation type, arc-related partial discharge is used, making... The moment of arc initiation. For arc event indication, This indicates that it has been confirmed as an arc light. For the arc light pre-correlation time window, when and At that time, i.e., before the arc light Significant partial discharge was observed within a short period of time, indicating insulation degradation and triggering a trip. S3-2. For non-insulation deterioration type, arc-correlated overcurrent is used, and the sliding window length is determined based on the sampling rate and system power frequency. Let the discrete values ​​of the current sampling sequence be... Mean within the window ,in The sample number corresponding to the current calculation time, and the effective value of the current. ,when and When an arc occurs and the current exceeds the overcurrent threshold within the associated window, it is judged as a non-insulation degradation type and trips. This is the overcurrent setting value. For overcurrent related windows.

[0010] According to the above technical solution, in step S4, the process of establishing the nonlinear correspondence is as follows: [The process involves] converting the discrete sequence... Aligned with the time axis, the continuous-time function is obtained by reconstructing the interpolation. The time change rate of the effective value of the current For switchgear of the same model, its and Based on similarity, paired sequences are extracted using a uniform statistical step size Δ. Let the fit be polynomial form of degree ,in The regression coefficients are obtained by fitting historical data to the same model of switchgear. and The correspondence is determined by converting the sample input into a multinomial feature matrix, using the least squares method to solve the parameter vector to obtain the magnitude of the regression coefficients, thereby determining the nonlinear correspondence.

[0011] According to the above technical solution, in step S5, the method for calculating the deviation between the actual change and the expected value is as follows: at a certain moment... In, according to S2 The calculation formula calculates the actual rate of change. The corresponding value is calculated based on the time-varying rate of change of the actual measured effective current value. Then, the expected rate of change is calculated using the nonlinear correspondence of S4. , when the residual When the threshold is exceeded, The residual threshold is used to determine the following: If a continuous positive bias occurs, the increase in the comprehensive partial discharge index is faster than the rate at which the load can explain it, which is determined to be one of the following: defect activation, surface creepage initiation, or partial discharge channel opening. If a continuous negative bias occurs, the comprehensive partial discharge index does not increase accordingly when the load increases, which is determined to be nonlinear behavior such as blocked energy transmission path, abnormal sensing attenuation, or temporary passivation and reactivation of the partial discharge channel.

[0012] According to the above technical solution, in step S5, when the residual continuously exceeds the threshold, the arc-related partial discharge determination is enhanced to relax the correlation threshold. And widen the arc light pre-correlation time window It synchronously records the window data before and after the triggering of key quantities such as arcing, partial discharge, and current, forming event segments, and reports the event, level, and suggested maintenance conclusions through the human-machine interface and communication port.

[0013] Compared with existing technologies, the beneficial effects achieved by this invention are as follows: This invention deeply integrates three criteria—partial discharge, arcing, and overcurrent—to construct a system that can provide early warning and precise protection for insulation degradation faults, as well as reliable backup protection for non-insulation degradation sudden faults. It achieves panoramic monitoring of two major categories of faults: slow insulation degradation and sudden short circuits. Through time-series correlation and multi-source cross-confirmation, it avoids false tripping caused by single triggering. It upgrades overcurrent from a single amplitude threshold to a multi-dimensional electrical quantity replacement trigger, allowing the system to operate when arcing is combined with a weak but obviously abnormal current, avoiding failure to operate due to insufficient current. Attached Figure Description

[0014] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a schematic diagram of the overall modular structure of the present invention. Detailed Implementation

[0015] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] Please see Figure 1 This invention provides a technical solution: a high-voltage switchgear intelligent monitoring and fault adaptive protection system based on multi-source signal fusion, comprising a multi-source sensing module, a state assessment module, and an adaptive protection module. The multi-source sensing module synchronously collects UHF, AE, arc, and primary circuit current, completes front-end conditioning and unified time reference, which serves as the basis for subsequent correlation and calculation. The state assessment module completes classification based on dual thresholds and trends, and establishes and calls the nonlinear relationship between partial discharge and overcurrent according to the same model baseline, calculates residuals and persistence online, and outputs level, high-risk flag and auxiliary criteria. The adaptive protection module executes arc correlation partial discharge and arc overcurrent fusion triggering, trips when the fusion criteria are met, and enhances the sensitivity of the correlation threshold and pre-window when the residual continuously exceeds the threshold. The multi-source sensing module includes a multi-source sensing unit and a fusion judgment module. The multi-source sensing unit consists of a UHF sensor, an ultrasonic sensor, an arc light sensor, and a current sampling module, used to collect four signals: UHF, AE, arc light, and current. The fusion judgment module performs PRPD spectrum analysis on the partial discharge signal to extract the characteristics of discharge quantity and discharge frequency; and calculates the effective value of the current signal. The condition assessment module includes a dual-threshold trend enhancement grading module, a nonlinear baseline modeling module, and a residual determination module. The dual-threshold trend enhancement grading module is used to output normal, first-level, and second-level grades with low and high threshold superimposed trends and persistence rules, focusing on sensitive identification of continuous degradation. The nonlinear baseline modeling module establishes and calls the nonlinear mapping of partial discharge and overcurrent for each model, providing expected changes in load interpretability. The residual determination module calculates whether the residuals of the actual and expected changes continue to deviate. The adaptive protection module includes an associated threshold adjustment module, a trip control module, and an event packaging and reporting module. The associated threshold adjustment module lowers the partial discharge associated threshold by a coefficient when nonlinear inconsistency occurs. The event packaging and reporting module sends various data as structured evidence along with the level and action results. The trip control module is used to trip when a fault occurs. The system works as follows: S1. Simultaneously acquire four signals: UHF, AE, arc, and current, and complete front-end analog and digital conditioning. Generate pulse sequences and amplitude statistics for the partial discharge channel that can be used for PRPD analysis. Calculate the effective value and over-limit flag for the current. Align the timestamps of each channel and establish a consistent time reference for subsequent correlation of partial discharge, arc, and overcurrent. The current is the operating current of the primary circuit. S2. Based on the partial discharge characteristic value and trend, the system is classified into normal, first-level early warning and second-level alarm levels, respectively, and corresponding prompts for attention, suggestions for enhanced maintenance and local audible and visual reporting are provided. The threshold adopts a dual-threshold structure and combines trend to enhance the sensitivity to continuous deterioration. When the second-level alarm state is entered, the system marks the insulation as high risk, providing a priori conditions for the sensitive switching of the protection logic. S3. The triggering criteria adopt configurable logic of arc-related partial discharge and arc-related overcurrent to achieve both prevention of false tripping and prevention of failure to trip. The insulation deterioration type fault is triggered by the arc-related partial discharge signal to suppress false tripping caused by pure optical interference. The non-insulation deterioration type sudden short circuit is backed by arc-related + overcurrent to ensure that the fault is cleared without missing detection. When the fusion criteria are met, a trip command is immediately output to clear the fault. S4. For the same type of switchgear, based on the collected partial discharge signal and current RMS value, establish a nonlinear correspondence between the time change rate of the partial discharge comprehensive index and the time change rate of the current RMS value, and form a baseline for this type of switchgear. S5. During operation, the two are synchronously calculated at the same time step and substituted into the baseline relationship to obtain the expected change. If the deviation between the actual change and the expected value continues to exceed the threshold, it is judged as an abnormal sign that is inconsistent with simple load fluctuations, and is used as an auxiliary criterion for alarm and protection triggering. In S2, the alarm classification is specifically as follows: The device provides amplitude statistics for the partial discharge channels UHF and AE. With pulse statistics Threshold discrimination is performed in a sliding window. The amplitudes of AE and UHF are taken respectively. With pulse counting ,calculate , ,in , These are the normal amplitude threshold and the high amplitude threshold, respectively. , These represent the normal and high thresholds for pulse counting, and the comprehensive partial discharge index. , For the linear contribution of the pulse, The amplification factor is used to classify the system into three levels: normal, first-level warning, and second-level alarm. A value below the preset alarm threshold is marked as normal; a value exceeding the lower threshold is marked as... However, a Level 1 warning is issued when the development trend is stable; otherwise, a high threshold is exceeded. And when it shows a continuous increasing trend, that is A level two alarm is triggered when the value increases within multiple consecutive sliding windows. The classification criterion is based on a coupling of three elements: low / high threshold, time trend, and persistence. The fusion index S(t) simultaneously covers both amplitude and impulse activity, making it mathematically clear to distinguish between occasional spikes and persistent degradation. This approach, which combines amplitude threshold, growth slope, and persistence ratio within the same criterion system, is more stable in identifying early degradation than existing thresholds that only consider amplitude or only consider frequency.

[0017] In S3, the triggering criteria are as follows: S3-1. For insulation degradation type, arc-related partial discharge is used, making... The moment of arc initiation. For arc event indication, This indicates that it has been confirmed as an arc light. For the arc light pre-correlation time window, when and At that time, i.e., before the arc light Significant partial discharge was observed within a short period of time, indicating insulation degradation and triggering a trip. S3-2. For non-insulation deterioration type, arc-correlated overcurrent is used, and the sliding window length is determined based on the sampling rate and system power frequency. Let the discrete values ​​of the current sampling sequence be... Mean within the window ,in The sample number corresponding to the current calculation time, and the effective value of the current. ,when and When an arc occurs and the current exceeds the overcurrent threshold within the associated window, it is judged as a non-insulation degradation type and trips. This is the overcurrent setting value. For overcurrent-related windows; By introducing a pre-correlation window Tpre with the arc light as the time anchor point, a verifiable partial discharge precursor is required within a short window before the arc light, forming a causal constraint on the arc light-related partial discharge. At the same time, the arc light overcurrent is retained in parallel as a backup for non-insulating faults. This structure of causal correlation + parallel backup not only suppresses false tripping due to pure optical interference, but also avoids failure to trip due to high impedance faults. It is fundamentally different from the traditional parallel dual criteria of arc light + overcurrent.

[0018] In S4, the process of establishing the nonlinear correspondence is as follows: The discrete sequence... Aligned with the time axis, the continuous-time function is obtained by reconstructing the interpolation. The time change rate of the effective value of the current For switchgear of the same model, its and Based on similarity, paired sequences are extracted using a uniform statistical step size Δ. Let the fit be polynomial form of degree ,in The regression coefficients are obtained by fitting historical data to the same model of switchgear. and The correspondence is determined by converting the sample input into a multinomial feature matrix, using the least squares method to solve the parameter vector to obtain the magnitude of the regression coefficients, thereby determining the nonlinear correspondence. Simultaneously with the triggering action, multiple channels such as arc, partial discharge, and current are synchronously recorded within a millisecond window before and after the triggering point according to a unified time base. Additionally, derived quantities such as ∇S / ∇I, expected ∇S, and residuals are packaged and uploaded as structured evidence, integrating protection actions with traceable diagnosis. Unlike existing practices that only store the original waveform or only report status codes, this design, which solidifies the key intermediate quantities in the criterion link, significantly improves post-event interpretability and operational decision-making efficiency.

[0019] In S5, the method for calculating the deviation between the actual change and the expected value is as follows: at a certain moment... In, according to S2 The calculation formula calculates the actual rate of change. The corresponding value is calculated based on the time-varying rate of change of the actual measured effective current value. Then, the expected rate of change is calculated using the nonlinear correspondence of S4. , when the residual When the threshold is exceeded, The residual threshold is used. If there is a continuous positive bias, the rise of the partial discharge comprehensive index is faster than the rate that the load can explain. This is judged to be one of the following: defect activation, surface creepage initiation, or partial discharge channel opening. If there is a continuous negative bias, the partial discharge comprehensive index does not rise accordingly when the load increases. This is judged to be nonlinear behavior such as blocked energy transmission path, abnormal sensing attenuation, or temporary passivation and reactivation of partial discharge channel. A nonlinear baseline of ∇S–∇I is established for the same model. The difference between the expected change obtained from online prediction and the measured change is used to form the residual. The dual thresholds of exceeding the limit and persistence are used as auxiliary criteria for inconsistency with the load. Compared with the static threshold or linear correlation test commonly used in the industry, this approach models the interpretable changes of the load first and then strips them out. Only the remaining deviations enter the alarm link, which reduces false alarms and can expose abnormal evolution in advance.

[0020] In S5, when the residual continuously exceeds the threshold, the arc-related partial discharge determination is enhanced by relaxing the correlation threshold. And widen the arc light pre-correlation time window It synchronously records the window data before and after the triggering of key quantities such as arcing, partial discharge, and current, forming event segments, and reports the event, level, and suggested maintenance conclusions through the human-machine interface and communication port.

[0021] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0022] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A high-voltage switchgear intelligent monitoring and fault adaptive protection system with multi-source signal fusion, characterized in that: It includes a multi-source sensing module, a state assessment module, and an adaptive protection module. The multi-source sensing module synchronously collects UHF, AE, arc light, and primary circuit current to complete front-end conditioning and unify the time reference, which serves as the basis for subsequent correlation and calculation. The state assessment module completes the classification based on dual thresholds and trends, and establishes and calls the nonlinear relationship between partial discharge and overcurrent according to the same model baseline, calculates residuals and persistence online, and outputs the level, high-risk indicator, and auxiliary criteria. The adaptive protection module performs arc light-related partial discharge and arc light overcurrent fusion triggering, trips when the fusion criteria are met, and enhances the sensitivity of the correlation threshold and pre-window when the residuals continue to exceed the threshold.

2. The intelligent monitoring and fault adaptive protection system for high-voltage switchgear based on multi-source signal fusion as described in claim 1, characterized in that: The multi-source sensing module includes a multi-source sensing unit and a fusion judgment module. The multi-source sensing unit consists of a UHF sensor, an ultrasonic sensor, an arc light sensor, and a current sampling module, which is used to collect four signals: UHF, AE, arc light, and current. The fusion judgment module performs PRPD spectrum analysis on the partial discharge signal to extract the characteristics of discharge quantity and discharge frequency. Calculate the effective value of the current signal; The state assessment module includes a dual-threshold trend enhancement grading module, a nonlinear baseline modeling module, and a residual determination module. The dual-threshold trend enhancement grading module is used to output normal, first-level, and second-level grades with low and high threshold superimposed trends and persistence rules, focusing on sensitive identification of continuous degradation. The nonlinear baseline modeling module establishes and calls nonlinear mappings of partial discharge and overcurrent for each model, providing expected changes in load interpretability. The residual determination module calculates whether the residual determination of actual and expected changes continues to deviate. The adaptive protection module includes an association threshold adjustment module, a trip control module, and an event packaging and reporting module. The association threshold adjustment module lowers the partial discharge association threshold by a coefficient when nonlinear inconsistency occurs. The event packaging and reporting module sends various data as structured evidence along with the level and action results. The trip control module is used to trip when a fault occurs.

3. The intelligent monitoring and fault adaptive protection system for high-voltage switchgear based on multi-source signal fusion according to claim 2, characterized in that: The system works as follows: S1. Simultaneously acquire four signals: UHF, AE, arc, and current, and complete front-end analog and digital conditioning. Generate pulse sequences and amplitude statistics for the partial discharge channel that can be used for PRPD analysis. Calculate the effective value and over-limit flag for the current. Align the timestamps of each channel and establish a consistent time reference for subsequent correlation of partial discharge, arc, and overcurrent. The current is the operating current of the primary circuit. S2. Based on the partial discharge characteristic value and trend, the system is classified into normal, first-level early warning and second-level alarm levels, respectively, and corresponding prompts for attention, suggestions for enhanced maintenance and local audible and visual reporting are provided. The threshold adopts a dual-threshold structure and combines trend to enhance the sensitivity to continuous deterioration. When the second-level alarm state is entered, the system marks the insulation as high risk, providing a priori conditions for the sensitive switching of the protection logic. S3. The triggering criteria adopt configurable logic of arc-related partial discharge and arc-related overcurrent to achieve both prevention of false tripping and prevention of failure to trip. The insulation deterioration type fault is triggered by the arc-related partial discharge signal to suppress false tripping caused by pure optical interference. The non-insulation deterioration type sudden short circuit is backed by arc-related + overcurrent to ensure that the fault is cleared without missing detection. When the fusion criteria are met, a trip command is immediately output to clear the fault. S4. For the same type of switchgear, based on the collected partial discharge signal and current RMS value, establish a nonlinear correspondence between the time change rate of the partial discharge comprehensive index and the time change rate of the current RMS value, and form a baseline for this type of switchgear. S5. During operation, both are calculated synchronously at the same time step and substituted into the baseline relationship to obtain the expected change. If the deviation between the actual change and the expected value continues to exceed the threshold, it is judged as an abnormal sign that is inconsistent with simple load fluctuations, and is used as an auxiliary criterion for alarm and protection triggering.

4. The intelligent monitoring and fault adaptive protection system for high-voltage switchgear based on multi-source signal fusion according to claim 3, characterized in that: In S2, the alarm classification is specifically as follows: The device provides amplitude statistics for the partial discharge channels UHF and AE. With pulse statistics Threshold discrimination is performed in a sliding window. The amplitudes of AE and UHF are taken respectively. With pulse counting ,calculate , ,in , These are the normal amplitude threshold and the high amplitude threshold, respectively. , These represent the normal and high thresholds for pulse counting, and the comprehensive partial discharge index. , For the linear contribution of the pulse, The amplification factor is used to classify the system into three levels: normal, first-level warning, and second-level alarm. A value below the preset alarm threshold is marked as normal; a value exceeding the lower threshold is marked as... However, a Level 1 warning is issued when the development trend is stable; otherwise, a high threshold is exceeded. And when it shows a continuous increasing trend, that is The system triggers a level 2 alarm when the value increases within multiple consecutive sliding windows.

5. The intelligent monitoring and fault adaptive protection system for high-voltage switchgear based on multi-source signal fusion according to claim 4, characterized in that: In S3, the triggering criteria are specifically as follows: S3-1. For insulation degradation type, arc-related partial discharge is used, making... The moment of arc initiation. For arc event indication, This indicates that it has been confirmed as an arc light. For the arc light pre-correlation time window, when and At that time, i.e., before the arc light Significant partial discharge was observed within a short period of time, indicating insulation degradation and triggering a trip. 3-2. For non-insulation deterioration type, arc-correlated overcurrent is used, and the sliding window length is determined based on the sampling rate and system power frequency. Let the discrete values ​​of the current sampling sequence be... Mean within the window ,in The sample number corresponding to the current calculation time, and the effective value of the current. ,when and When an arc occurs and the current exceeds the overcurrent threshold within the associated window, it is judged as a non-insulation degradation type and trips. This is the overcurrent setting value. For overcurrent related windows.

6. The intelligent monitoring and fault adaptive protection system for high-voltage switchgear based on multi-source signal fusion according to claim 5, characterized in that: In S4, the process of establishing the nonlinear correspondence is as follows: The discrete sequence... Aligned with the time axis, the continuous-time function is obtained by reconstructing the interpolation. The time change rate of the effective value of the current For switchgear of the same model, its and Based on similarity, paired sequences are extracted using a uniform statistical step size Δ. Let the fit be polynomial form of degree ,in The regression coefficients are obtained by fitting historical data to the same model of switchgear. and The correspondence is determined by converting the sample input into a multinomial feature matrix, using the least squares method to solve the parameter vector to obtain the magnitude of the regression coefficients, thereby determining the nonlinear correspondence.

7. The intelligent monitoring and fault adaptive protection system for high-voltage switchgear based on multi-source signal fusion according to claim 6, characterized in that: In S5, the method for calculating the deviation between the actual change and the expected value is as follows: at a certain moment... In, according to S2 The calculation formula calculates the actual rate of change. The corresponding value is calculated based on the time-varying rate of change of the actual measured effective current value. Then, the expected rate of change is calculated using the nonlinear correspondence of S4. , when the residual When the threshold is exceeded, The residual threshold is used to determine the following: If a continuous positive bias occurs, the increase in the comprehensive partial discharge index is faster than the rate at which the load can explain it, which is determined to be one of the following: defect activation, surface creepage initiation, or partial discharge channel opening. If a continuous negative bias occurs, the comprehensive partial discharge index does not increase accordingly when the load increases, which is determined to be nonlinear behavior such as blocked energy transmission path, abnormal sensing attenuation, or temporary passivation and reactivation of the partial discharge channel.

8. The intelligent monitoring and fault adaptive protection system for high-voltage switchgear based on multi-source signal fusion according to claim 7, characterized in that: In step S5, when the residual continuously exceeds the threshold, the arc-related partial discharge determination is enhanced to relax the correlation threshold. And widen the arc light pre-correlation time window It synchronously records the window data before and after the triggering of key quantities such as arcing, partial discharge, and current, forming event segments, and reports the event, level, and suggested maintenance conclusions through the human-machine interface and communication port.

Images