Font Size: a A A

Research On Pattern Recognition Technology Of Water Pipeline Vibration Signal Based On Support Vector Machine

Posted on:2022-12-06Degree:MasterType:Thesis
Country:ChinaCandidate:H Q ZhuFull Text:PDF
GTID:2492306743473064Subject:Control Engineering
Abstract/Summary:PDF Full Text Request
This paper built a water pipeline early warning system based on phase-sensitive optical time domain reflectometry(Φ-OTDR)in order to ensure the safe and efficient operation of water pipelines and realize the real-time detection of water pipeline status.The designed algorithm realized the pipeline vibration signal processing and pattern recognition and reduced the false alarm rate of the system.The contents of this paper are as follows:(1)The vibration signal of PZT piezoelectric ceramics collected in the laboratory was denoised.The equalization optimization algorithm was used to optimize the parameters of the variational mode decomposition algorithm and decompose the original signal.The components with a large degree of correlation with the original signal were selected and retained,and the components with a small degree of correlation were denoised for the second time with the singular value decomposition method and reconstructed,and the final reconstructed signal can greatly remove noise.After treatment with the algorithm in this paper,the signal-to-noise ratio of the signal can reach 13 d B,which lays the foundation for the pattern recognition in the next step.(2)The pattern recognition analysis was carried out on the denoised signal,and the features of the signal was extracted from the time domain,frequency domain and space domain with multi-feature fusion method.Then it proposes to use an improved two-level support vector machine algorithm to classify the five vibration events,such as human touch,manual excavation and mechanical excavation.The overall recognition efficiency of five vibration events is 95.46%,and the recognition accuracy of the system is high,which can meet the requirement for engineering application.(3)The Φ-OTDR software/hardware system was built.The parameters of the system devices were analyzed in detail,and the devices were selected according to the technical indicators of the Φ-OTDR system.Finally,the interface of the upper computer was designed,and the real-time data display,vibration point processing and pattern recognition functions were realized.In this paper,the real-time detection of the pipeline state was achieved by applying the distributed optical fiber sensing technology to the detection of pipeline vibration events.Practical tests show that the algorithm proposed in this paper can realize the precise location and pattern recognition of vibration events near the pipeline,and therefore can be further applied in engineering.
Keywords/Search Tags:Phase-sensitive optical time domain reflectometry, Signal denoising, feature extraction, pattern recognition, Pipeline safety inspection
PDF Full Text Request
Related items