CN-122020440-A - Predictive analysis system based on optical fiber vibration detection

Abstract

The invention discloses a prediction analysis system based on optical fiber vibration detection, and relates to the technical field of optical fiber vibration detection. According to the prediction analysis system based on the optical fiber vibration detection, the double optical fibers are combined with the self-adaptive noise elimination algorithm, so that equipment vibration and environmental noise can be effectively separated, signal quality is improved, a dynamic physical map generator is used for constructing a dynamic transfer function, dynamic factors such as equipment aging and the like are considered, environmental influence is compensated through the environmental attenuation function, prediction precision and anti-interference capability are improved, and a fault tracing module can accurately position a fault source by utilizing jacobian matrix regularization calculation. The migration learning engine can realize rapid modeling of new equipment based on physical mechanism classification and constraint conditions, reduce implementation cost and time, and enhance system generality and expandability.

Inventors

• LIU HAO
• YANG JIAN
• WU LIFANG
• WANG HAIPENG
• GUO TAOTAO
• WANG XIAOBIN
• SUN YE
• HE ZHIBIN
• ZHANG SIYANG

Assignees

• 内蒙古电力(集团)有限责任公司呼和浩特供电分公司

Dates

Publication Date

20260512

Application Date

20251208

Claims (6)

1. 1. The prediction analysis system based on the optical fiber vibration detection is characterized by comprising an optical fiber sensing module, a causal decoupling processor, a dynamic physical map generator and a fault tracing module: the optical fiber sensing module comprises a main sensing optical fiber and an environment reference optical fiber; the causal decoupling processor is configured to decompose the original optical signal into: the dynamic physical map generator outputs a predicted vibration: Inputting an environmental vector ; The fault tracing module reversely pushes abnormal parameters through the frequency offset: 。
2. 2. A predictive analysis system based on optical fiber vibration detection as claimed in claim 1, wherein: the vibration transfer function Constructed from the following model: 。
3. 3. A predictive analysis system based on optical fiber vibration detection as claimed in claim 1, wherein: the causal decoupling processor performs a noise cancellation algorithm: ; Optimum coefficient By solving: ; Updating when the equipment is shut down: 。
4. 4. A predictive analysis system based on optical fiber vibration detection as claimed in claim 1, wherein: The fault tracing module adopts regularization calculation: When (when) And judging that the j parameter is abnormal.
5. 5. A predictive analysis system based on optical fiber vibration detection as claimed in claim 1, wherein: the environmental decay function is defined as: 。
6. 6. the prediction analysis system based on optical fiber vibration detection according to claim 1, wherein the system further comprises a migration learning engine, and the constraint new equipment meets the following conditions: 。

Description

Predictive analysis system based on optical fiber vibration detection Technical Field The invention relates to the technical field of optical fiber vibration detection, in particular to a predictive analysis system based on optical fiber vibration detection. Background The vibration of the optical fiber refers to a phenomenon that when external factors act on the optical fiber, the optical fiber is slightly deformed and vibrated, and optical parameters such as phase, amplitude, frequency and the like of transmitted light in the optical fiber are changed. These external factors may be mechanical vibrations, pressure changes, temperature fluctuations, etc. The optical fiber vibration detection has various important meanings. In the security field, the system can be used for perimeter precaution, such as optical fiber vibration sensing systems are paved around places like airports, military bases, prisons and the like, and once optical fiber vibration is caused by illegal invasion, the system can timely detect and send out an alarm, so that real-time monitoring and early warning are realized, and the safety of places is ensured. In the aspect of monitoring the infrastructure, for bridges, tunnels, large buildings and the like, the information such as stress, strain, vibration state and the like of the structure can be obtained by detecting the vibration of the optical fiber, hidden dangers such as structural damage, fatigue and the like can be found in time, a basis is provided for health evaluation and maintenance of the structure, and the safe and stable operation of the infrastructure is ensured. In the oil and gas industry, the monitoring device can be used for monitoring long-distance oil and gas pipelines, detecting whether the pipelines are damaged by a third party, affected by geological disasters or leaked, and the like, and guaranteeing the safe operation of the pipelines. In the railway transportation field, the vibration condition of the rail can be monitored, and the problems of abrasion and deformation of the rail, abnormal vibration during train operation and the like can be found in time, so that the running safety of the train is ensured. In the prior art, the traditional system usually only depends on a single optical fiber in a signal acquisition stage, and equipment vibration and environmental noise cannot be effectively distinguished, so that the signal-to-noise ratio of the signal is low, and equipment state information is difficult to accurately extract. In addition, modeling of equipment vibration in the prior art is too simplified, and a static transfer function is generally adopted, so that influences of dynamic factors such as equipment aging, operation time and the like on vibration characteristics cannot be reflected, and the deviation of a prediction result is larger. In addition, the interference of environmental factors (such as temperature, humidity and wind speed) on vibration signals is not fully processed, and an effective environmental attenuation compensation mechanism is lacked, so that the reliability of the system is further reduced. In the aspect of fault diagnosis, the prior art relies on an empirical threshold or simple spectrum analysis, lacks fine solution to physical parameter deviation, and is difficult to accurately locate a fault source. Disclosure of Invention Aiming at the defects in the prior art, the invention provides a predictive analysis system based on optical fiber vibration detection, which comprises an optical fiber sensing module, a causal decoupling processor, a dynamic physical map generator and a fault tracing module: the optical fiber sensing module comprises a main sensing optical fiber and an environment reference optical fiber; the causal decoupling processor is configured to decompose the original optical signal into: ; Wherein: Is the vibration transfer function of the i-th equipment, and is input into a physical state vector (Including rotational speed, temperature, pressure),Is the vibration coupling coefficient of the device,Is the gain factor of the ambient noise and,Is an ambient noise basis function; the dynamic physical map generator outputs a predicted vibration: wherein Input of an environmental vector as an environmental decay functionTemperature, humidity, wind speed); the fault tracing module reversely pushes abnormal parameters through the frequency offset: Wherein: Is the measured spectral shift vector and, Is a jacobian matrix that is a matrix of jacobian,Is the deviation of the physical parameter. Preferably, the vibration transfer functionConstructed from the following model: and wherein: Is a standard device transfer function that is used to determine the transfer characteristics of the device, Is the amount of the ageing correction,Is the run-time dependent decay factor. Preferably, the causal decoupling processor performs a noise cancellation algorithm: Optimal coefficient By solving: . Updating when the equipment i