Research And Implementation Of Optical Fiber Vibration Signal Recognition Based On Intelligent Learning

With the rapid development of economy and science and technology in China,the demand for oil and gas resources has increased,and pipeline transportation of oil and gas has become the main mode of transportation.Although our country has invested a lot of financial and material resources in oil and gas pipeline monitoring,it still can not meet the accuracy and real-time requirements for inspection and identification.When oil and gas pipelines are damaged,greater losses can be avoided if the exact location of the accident and the way in which the pipeline was damaged can be accurately obtained,and staff can be sent to the accident site to inspect and respond to the damage.To ensure the security of the property of the State and the people.In view of this situation,compared with the traditional detection and recognition technology,optical fiber early warning technology with better performance began to develop rapidly.In this paper,intelligent learning is introduced into the identification of optical fiber vibration signals.By extracting the pitch frequency,energy duty ratio and duty ratio based on wavelet packet decomposition of various types of vibration signals collected from the field,the characteristics of the pitch frequency,the energy duty ratio and the duty cycle of the vibration signals collected in the field are extracted.The BP neural network is used to train these features,and a neural network model which can distinguish many different vibration signals is obtained.Among them,this paper uses BP neural network pre-training and cross-validation method to select the number of hidden layer nodes to optimize the vibration signal neural network model.Compared with the traditional vibration signal recognition technology,the network model obtained by intelligent learning and training effectively improves the accuracy and reliability of vibration source signal identification.In this paper,based on the optical fiber early warning detection and recognition system,the characteristics and performance of the system are analyzed,and a set of optical fiber early warning detection and recognition system software is designed and implemented according to the practical application and customer requirements.The software is divided into six modules,namely:software login verification management module,map calibration and map display module,detection and recognition module,parameter setting and visual display module,optical fiber refraction detection module,file transfer module.Through these modules,the detection and recognition results of the system can be better displayed to the user.