CN-122026598-A - High-voltage equipment fault intelligent prediction system based on arc light monitoring and optical fiber temperature measurement

CN122026598ACN 122026598 ACN122026598 ACN 122026598ACN-122026598-A

Abstract

The invention relates to the technical field of power system automation and high-voltage equipment state monitoring, and discloses a high-voltage equipment fault intelligent prediction system based on arc light monitoring and optical fiber temperature measurement. The invention dynamically adjusts a fluorescent optical fiber temperature sensor sampling strategy according to ultraviolet arc light intensity through a time domain cooperative controller, realizes multi-dimensional physical quantity time domain association capture by matching with an annular buffer zone, carries out cryptographic transmission by utilizing a WAPI gateway, carries out logic verification on a heterogeneous sensor by utilizing electric energy quality data based on a physical mutual exclusivity principle, and carries out thermal inertia compensation prediction by adopting a RCNN model. The problem of mismatch between transient arc light and steady-state temperature signal sampling is solved, false alarm of faults of a sensor body is effectively eliminated, and heat conduction hysteresis is overcome to realize advanced accurate early warning.

Inventors

LIN LIANGPEI
CHENG GUANJI
LIU SONGLIN
GAO FENG
CHEN MO
WANG JIAQI
ZHANG YUANXIA
ZHANG DAWU

Assignees

中国南方电网有限责任公司超高压输电公司广州局

Dates

Publication Date: 20260512
Application Date: 20260109

Claims (10)

1. High-voltage equipment trouble intelligent prediction system based on arc light monitoring and optic fibre temperature measurement, its characterized in that includes: The multi-source physical sensing module is deployed on a high-voltage equipment monitoring site and is used for monitoring physical signals and outputting monitoring data, and the multi-source physical sensing module comprises an ultraviolet light fiber arc light sensing unit, a fluorescent fiber temperature sensing unit and an electric energy quality sensing unit; The time domain cooperative acquisition control module is configured with cooperative logic and is used for dynamically adjusting the data acquisition strategy of the fluorescent optical fiber temperature sensing unit according to the monitoring state of the ultraviolet optical fiber arc light sensing unit; the secure transmission and gateway module comprises a WAPI secure gateway and is used for receiving the monitoring data, executing encryption processing based on a national encryption algorithm and transmitting the encrypted data through a wireless network; The system comprises a server, a heterogeneous data self-checking module, a data processing module and a data processing module, wherein the heterogeneous data self-checking module is arranged at the server and is used for receiving and analyzing encrypted data, and the heterogeneous data self-checking module is also used for respectively carrying out logic checking on fault states of the ultraviolet light fiber arc light sensing unit and the fluorescent light fiber temperature sensing unit by utilizing the data of the electric energy quality sensing unit based on a physical mutual exclusivity principle, removing abnormal data and outputting checked data; And the intelligent diagnosis and interaction module is used for receiving the data passing the verification and constructing a thermal inertia compensation prediction model, and outputting a device fault prediction result based on the time domain correlation of the optical signal and the thermal signal.
2. The intelligent prediction system for fault of high-voltage equipment based on arc light monitoring and optical fiber temperature measurement according to claim 1, wherein the dynamic adjustment of the data acquisition strategy of the fluorescent optical fiber temperature sensing unit by the time domain cooperative acquisition control module comprises: presetting a weak discharge trigger threshold; when the signal intensity of the ultraviolet light fiber arc light sensing unit is monitored to not exceed the weak discharge trigger threshold, controlling the fluorescent light fiber temperature sensing unit to run in a low-frequency mode, and only collecting steady-state heat balance data; when the signal intensity of the ultraviolet light fiber arc light sensing unit is monitored to exceed the weak discharge triggering threshold, a control instruction is sent to the fluorescent light fiber temperature sensing unit, the fluorescent light fiber arc light sensing unit is switched to a high-frequency mode to operate, high-frequency sampling is kept in a preset high-frequency acquisition time window, and transient temperature rise data are captured.
3. The arc monitoring and fiber temperature measurement based high voltage equipment fault intelligent prediction system according to claim 2, wherein the time domain cooperative acquisition control module is internally provided with a ring buffer for performing thermal inertia data capture, and the system comprises: When the fluorescent optical fiber temperature sensing unit operates in a low-frequency mode, continuously writing data into the annular buffer area and circularly covering the annular buffer area; when switching to a high-frequency mode, locking current stored data in the annular buffer zone to serve as preposed background data, and synchronously recording transient light pulse data at a trigger moment and high-frequency temperature rise curve data in the high-frequency acquisition time window; And splicing the front background data, the transient light pulse data and the high-frequency temperature rise curve data according to a unified time axis to form a fault event data packet containing the whole process from the light energy injection stage to the heat energy dissipation stage.
4. The intelligent prediction system for fault of high-voltage equipment based on arc light monitoring and optical fiber temperature measurement according to claim 1, wherein the heterogeneous data self-checking module performs logic checking on the fault state of the fluorescent optical fiber temperature sensing unit by using the data of the power quality sensing unit, wherein the logic checking comprises: calculating the real-time temperature change rate of the data uploaded by the fluorescent optical fiber temperature sensing unit; acquiring a load current effective value uploaded by the power quality sensing unit and an arc intensity integral value uploaded by the ultraviolet light fiber arc sensing unit; and when the real-time temperature change rate is detected to exceed a preset temperature rise rate physical limit threshold, and the effective value of the load current at the same moment is lower than a preset overload threshold and the arc intensity integral value is lower than a preset arc discharge threshold, judging that the probe of the fluorescent optical fiber temperature sensing unit falls off or a circuit fault exists.
5. The intelligent prediction system for fault of high-voltage equipment based on arc light monitoring and optical fiber temperature measurement according to claim 1, wherein the heterogeneous data self-checking module performs logic checking on the fault state of the ultraviolet optical fiber arc light sensing unit by using the data of the power quality sensing unit, wherein the logic checking comprises: extracting the current total harmonic distortion rate and high-frequency pulse count of the data acquired by the power quality sensing unit; acquiring an ultraviolet intensity baseline value output by the ultraviolet fiber arc light sensing unit; And when the current total harmonic distortion rate is detected to exceed a preset electric discharge warning threshold or the high-frequency pulse count shows dense distribution characteristics, and the ultraviolet intensity baseline value in the same time window is lower than a preset background noise threshold, judging that the optical path of the ultraviolet fiber arc light sensing unit is polluted.
6. The intelligent prediction system for fault of high voltage equipment based on arc monitoring and optical fiber temperature measurement according to claim 1, wherein the intelligent diagnosis and interaction module comprises a data preprocessing unit, and the data preprocessing unit performs the following time domain alignment steps: establishing a unified time axis by taking the triggering time of the ultraviolet light fiber arc light sensing unit as a reference, and defining a unified time step; Generating virtual data points by adopting a cubic spline interpolation algorithm for the low-frequency temperature data before the trigger time, and mapping the low-frequency data to the uniform time step; For the pulse waveform collected by the ultraviolet light fiber arc light sensing unit, calculating the integral intensity in the unified time step as the light energy density characteristic; And performing normalization processing on the aligned multidimensional feature matrix to generate a time sequence tensor for model input.
7. The intelligent prediction system for high-voltage equipment faults based on arc light monitoring and optical fiber temperature measurement according to claim 6, wherein the thermal inertia compensation prediction model is constructed based on a recurrent convolutional neural network, and specifically comprises: the input layer of the thermal inertia compensation prediction model receives the time sequence tensor, and the cross-modal concurrency characteristics of current harmonic waves and ultraviolet light intensity are extracted through a one-dimensional convolutional neural network layer; the time sequence memory layer of the thermal inertia compensation prediction model adopts a long-period memory network or a gating circulation unit, and utilizes a gating mechanism to memorize the historical energy state so as to simulate the thermal inertia of a conductor; And the output layer of the thermal inertia compensation prediction model outputs an algorithm-corrected virtual core temperature, and the virtual core temperature is advanced from the actual measurement reading of the fluorescent optical fiber temperature sensing unit on a time axis and is used for representing the temperature of a real heating point in equipment.
8. The arc monitoring and fiber optic temperature measurement based high voltage equipment fault intelligent prediction system of claim 1, wherein the WAPI security gateway comprises a hardware security encryption chip and a local non-volatile memory, comprising: The SM4 block cipher algorithm logic is solidified in the hardware security encryption chip and is used for performing linear speed encryption on acquired data by using a session key generated by negotiation; the WAPI security gateway is provided with a broken chain protection logic, when the wireless network interruption is detected, collected data is redirected and written into the local nonvolatile memory, and after the wireless network is recovered, the data in the local nonvolatile memory is read based on a time stamp index for encryption and retransmission.
9. The intelligent prediction system for high voltage equipment failure based on arc light monitoring and optical fiber temperature measurement according to claim 1, wherein the WAPI security gateway, a wireless network access point and an authentication service unit together form an authentication architecture, wherein the WAPI security gateway is used as an authentication service user unit and is configured to invoke an internally stored digital certificate to initiate an authentication request, and negotiate with the wireless network access point to generate a session key for data encryption after passing the certificate validity verification of the authentication service unit.
10. The arc monitoring and fiber optic thermometry based high voltage equipment failure intelligent prediction system of claim 7, wherein the intelligent diagnostic and interactive module is further configured with a digital twinning based visualization unit comprising: Calculating the space coordinates of the discharge points by using a space difference algorithm of the signal intensity of the multi-path ultraviolet light fiber arc light sensing unit; In the three-dimensional digital twin interface, mapping the virtual core temperature output by the thermal inertia compensation prediction model into a volume fog effect thermodynamic diagram around the equipment model, and rendering the calculated space coordinates of the discharge points into a halo special effect.

Description

High-voltage equipment fault intelligent prediction system based on arc light monitoring and optical fiber temperature measurement Technical Field The invention relates to the technical field of power system automation and high-voltage equipment state monitoring, in particular to an intelligent high-voltage equipment fault prediction system based on arc light monitoring and optical fiber temperature measurement. Background The operation reliability of the high-voltage switch cabinet is directly related to the safety and stability of the power grid. Defects such as poor contact, insulation aging and the like in the equipment often cause partial discharge, overheat and even arc short circuit accidents. Currently, state monitoring for switch cabinets mainly relies on fluorescent fiber temperature measurement techniques and ultraviolet arc monitoring techniques. Because of the insulation performance, the fluorescent optical fiber is often attached to the surface of a high-voltage contact for temperature monitoring, and an ultraviolet probe is used for capturing weak ultraviolet light signals generated during insulation breakdown. However, existing monitoring systems often employ independent decentralized architectures, and temperature, arc and electrical quantity monitoring lacks efficient synergy in time domain and logic. The dynamic characteristics of different physical quantities have huge differences, the temperature belongs to a large inertia signal, low-frequency sampling is usually adopted, and the arc light belongs to microsecond transient signals. The steady state monitoring and transient state capturing cannot be considered by a normal rule frequency sampling strategy, transient temperature rise rate mutation is caused by a low-frequency sampling extremely easy missing fault, mass redundant data is generated by full-time high-frequency sampling, and the transmission and processing loads are exceeded. Second, single physical quantity discrimination lacks self-checking ability for the sensor's own health status. The fluorescent optical fiber probe is easy to measure the air temperature due to vibration falling, so that the passive high heat is caused, and the ultraviolet probe lens is easy to optically blind due to dust deposition and pollution. The existing scheme is difficult to accurately distinguish the hardware faults of the sensor from the faults of the equipment body, and false alarm or refusal is easy to cause. In addition, the surface temperature collected by the sensor has significant thermal conduction hysteresis relative to the conductor core temperature, affected by the thermal resistance effect of the insulating layer. The temperature overrun alarm is only dependent on actual measurement of the surface temperature, so that time delay caused by thermal inertia is difficult to correct, and early pre-prediction cannot be realized before irreversible thermal damage of the core component occurs. Disclosure of Invention Aiming at the defects of the prior art, the invention provides an intelligent high-voltage equipment fault prediction system based on arc light monitoring and optical fiber temperature measurement, which solves the problems of time domain mismatch of multi-source physical quantity, difficult self-detection of sensor faults and early warning lag caused by thermal inertia in the prior art. The intelligent high-voltage equipment fault prediction system based on arc light monitoring and optical fiber temperature measurement is realized by the following technical scheme. The system mainly comprises a multi-source physical sensing module, a time domain cooperative acquisition control module, a secure transmission and gateway module, a heterogeneous data self-checking module and an intelligent diagnosis and interaction module. The multi-source physical sensing module is deployed on a high-voltage equipment monitoring site and is used for monitoring multi-dimensional physical signals including an ultraviolet light fiber arc light sensing unit, a fluorescent light fiber temperature sensing unit and an electric energy quality sensing unit. The time domain collaborative acquisition control module is used for acquiring monitoring data and is configured with collaborative logic, and can dynamically adjust the data acquisition strategy of the fluorescent optical fiber temperature sensing unit according to the monitoring state of the ultraviolet optical fiber arc light sensing unit, so that the association capturing of different physical quantities on the time domain is realized. The secure transmission and gateway module comprises a WAPI secure gateway for receiving the monitoring data and executing encryption processing based on a national encryption algorithm, and transmitting the encrypted data through a wireless network. The heterogeneous data self-checking module is deployed at the server end, receives and analyzes the encrypted data, performs logic checking on fault states of the ultravi