Intelligent Detection Technology for Leakage Bag of Baghouse Based on Φ-OTDR Distributed Optical Fiber

Author:Liu Xu An

Supervisor:fang xiao dong zhang zhi rong


Degree Year:2019





In the process of industrial exhaust emissions,which contain particle pollutants,baghouse is often used for dust removal.The filter bag inside the baghouse will be damaged after being used for a period of time.It must be identified and located from a large number of filter bags inside the baghouse in time.Otherwise,dust particles will pass through the leak hole in the filter bag and be discharged into the air,resulting in air pollution.At present,the traditional identification and location methods have obvious shortcomings,such as high operating cost,long identification and location time,complex operation,and cannot achieve real-time online monitoring and positioning of filter bags.Distributed optical fiber vibration sensor system based on Phase-sensitive Optical Time Domain Reflectometer(Φ-OTDR)is widely used in many complex engineering applications because of its high detection accuracy,long sensing distance,high spatial resolution and real-time on-line location and monitoring of intrusion vibration events.It is widely used in many complex engineering application scenarios.In this paper,a Φ-OTDR distributed optical fiber sensing vibration system is used to monitor,locate and identify the damaged filter bags in the baghouse in real time.The research work in the following aspects is carried out concretely:Firstly,the theoretical knowledge related to Φ-OTDR sensing technology is introduced in detail,and a Φ-OTDR distributed optical fiber vibration sensor system is built based on the principle knowledge,and the technical parameters of the system are briefly introduced.Secondly,the laying method of transmission fiber in filter bag of bag filter is designed.In this paper,the sensing fiber is divided into several equal parts.The length of each aliquot is larger than the actual spatial resolution distance of theΦ-OTDR sensing system.A part of each equal part is folded and fixed under the condition that the fiber curvature is kept in a reasonable range.The folded optical fiber is suspended inside the filter bag,and the remaining optical fiber is reserved to ensure that the filter bag positioning program scans and locates each filter bag,a certain position of the optical fibers in the filter bag can be enough to locate the filter bag.Thirdly,on the experimental platform,the characteristics of the optical fiber vibration signal in the filter bag of the baghouse are analyzed.In this paper,the characteristics of vibration signals are analyzed in time domain,frequency domain and time-frequency domain.Different feature analysis algorithms are used in each analysis method,and the difference between the vibration signals of the good bag and the broken bag in the bag filter box is compared in detail.Then on the outfield bag filter platform,the vibration signal characteristics of the good bag and different types of broken bag inside the bag filter box were extracted and identified.According to the environmental protection requirements of dust emission in china,different types of damaged filter bags are manually manufactured and applied in the experiments.The vibration signal samples are decomposed based on the wavelet packet decomposition method,the energy spectrum and its energy entropy of each signal is calculated.Some sub-band energy and the energy entropy of the signal are extracted to form feature parameters,and back propagation neural network(BP)neural network algorithm is used to identify different types of filter bags.When the BP neural network is used for training and recognition testing,the samples used are from the same sample set.The purpose is to verify the feasibility of the algorithm in identifying and classifying the filter bags.The results show that the BP neural network classifier can effectively classify and identify different types of filter bags,the recognition rate is over 90%,and the recognition stability is good.Finally,the influence of the number of training samples on the recognition results is analyzed.Finally,on the outfield platform of bag filter,according to the practical requirements of engineering application,the software flow of detection,location and recognition of breakage filter bag is designed,and the feature extraction algorithm of the vibration signal in the filter bag and the training and recognition method of classifier samples are improved.The energy entropy and correlation coefficients are used to constitute the feature parameters,and a certain type of filter bag signal feature parameters are used as training samples.Then the trained recognizer is used to classify other different types of filter bags.The results show that the method can recognize and classify different types of filter bags very well,the recognition rate reaches more than 90%,and has good recognition stability.The real-time location and identification of the damaged filter bag in the bag filter are realized.This method has good engineering application value.