| Compared to other distributed sensing systems,the phase sensitive optical time domain reflectometer(Φ-OTDR)system based on heterodyne detection has simpler structure and higher sensitivity,which can meet the requirements of precise sensing in distributed vibration and acoustic sensing.With the development of national industrialization in recent years,raw material transportation belt conveyor has become an indispensable part of industrial production,and the safety guarantee of belt conveyor has accordingly become a hot topic of current research.In the future,the detection of belt conveyor fault signals will become a new application of Φ-OTDR system based on heterodyne detection.At the same time,it puts forward higher requirements for the performance of the system,such as signal-to-noise ratio(SNR),dynamic range and weak signal extraction.In this article,the detection of belt conveyor fault signals is the main research direction.A Φ-OTDR system based on heterodyne detection with high SNR and large dynamic range is proposed from three aspects: system principle,demodulation method and recognition method,and it achieves accurate detection and location of belt conveyor fault signals.The main work of this article is summarized as follows:1.For the improvement of the SNR of Φ-OTDR system based on heterodyne detection,a fading suppression algorithm based on optimal vector average is proposed.Firstly,the mathematical model relationship between the demodulated signal amplitude and the SNR of demodulated phase is established.Then based on this,a polarization state based fading suppression algorithm is proposed.Finally,a Φ-OTDR systems based on heterodyne detection is set up.Through experiments,the feasibility of the algorithm to suppress fading has been proved,and the algorithm can reduce system noise regardless of whether fading occurs.2.In order to solve the problem of dynamic range compressed in arctangent based In-phase Quadrature(IQ)digital demodulation,an optimized unwrapping algorithm based on wavelet decomposition and differential-compression-reduction is proposed.Firstly,the source of error in traditional unwrapping algorithms is analyzed,and the relationship between the maximum detectable amplitude of the algorithm and the sampling rate,as well as the signal frequency,is revealed.Then,corresponding algorithms are proposed to solve the unwrapping error caused by phase noise and phase difference values,respectively.Finally,experiments prove that,even when interference fading occurs,the signal can be restored normally by using this algorithm,and the dynamic range of the signal can be enlarged at the same time.3.Aiming at the vibration signals detected by distributed optical fiber sensing systems,a denoising algorithm combining singular value decomposition(SVD)and least mean square(LMS)adaptive filtering,as well as a feature extraction and fault location method based on power spectrum analysis and multi-peak detection,are proposed.The distributed sensing system proposed in this thesis is applied to the simulated fault scenarios of the belt conveyor.The experiments prove that the system can be used for fault vibration signal detection and location. |
