Search

Due to the low signal-to-noise ratio of raw data, the reliability of data and the accuracy of acoustic source localization are severely affected by fiber optic acoustic sensing technology. To address this issue, this study optimizes the wavelet thresholding method. Firstly, a novel thresholding function is proposed, which achieves denoising while preserving key information through shape adjustment factors. It combines the advantages of both hard and soft threshold functions and has high flexibility and controllability. Secondly, an adaptive threshold calculation method is introduced, utilizing an improved simulated annealing algorithm to optimize threshold selection, reducing the algorithm’s dependence on threshold parameter selection. Through simulation experiments, it has been verified that this research method effectively suppresses noise in the signal and improves data availability. Compared to the original methods, this approach significantly improves the signal-to-noise ratio and demonstrates robustness in simulated tests of real signals.

This paper proposes an improved adaptive noise-aided complete ensemble empirical mode decomposition (ICEEMDAN) method to address the issue of low signal-to-noise ratio in distributed optical fiber acoustic sensing systems. The proposed approach utilizes sample entropy and wavelet threshold denoising algorithm to extract valuable components from high noise components. The ICEEMDAN is applied to decompose the acquired signals, and sample entropy is calculated to identify the noisy components, which are then subjected to wavelet threshold denoising. Finally, the denoised components are reconstructed with the untreated intrinsic mode functions. Experimental results demonstrate that the denoising treatment significantly enhances the signal-to-noise ratio by 5.34 dB, reduces the mean square error by 0.014 8, and improves waveform similarity by 5.7%. Compared to other commonly used denoising methods, the proposed approach not only exhibits superior performance in terms of signal-to-noise ratio but also demonstrates better performance in mean square error and waveform similarity, thereby preserving useful signals more effectively.

Aiming at the problem that the PID control algorithm can not meet the requirement of tracking and docking of moving flange due to its slow response speed when controlling the LNG unloading arm, which has strong nonlinearity and time-varying system, the fuzzy control theory is introduced, and the fuzzy control algorithm based on experience and the fuzzy PID control algorithm which is combined with PID control are proposed and implemented respectively, which does not depend on the dynamic model of the LNG unloading arm. The two intelligent control algorithms are respectively applied to the LNG unloading arm of the test site. Taking a single joint as the test object, the performance test and the tracking test of the simulated moving target flange are carried out. The test show that the algorithm designed by introducing the fuzzy control theory has significantly improved the comprehensive response speed compared with the PID control algorithm, and the fuzzy PID control algorithm can fully meet the docking requirements of moving target flange in practical application because of its strong resistance to time delay.

Sequential pattern mining is always a very important branch of time series data mining. The pattern mining with visual means can be used to extract the knowledge of time series data more intuitively. Based on the research content, this paper analyzes the sequence pattern mining methods in different aspects and their combination with visualization technology. We further discuss and summarize the advantages of different visualization methods in discovering the potential patterns in time series data. Different systems and models have their unique information to show the focus. Compared with the characteristics of the model, the development and evolution of visualization technology for the discovery of potential patterns of time series data can be summarized. Finally, this paper discusses its development trend and how to play a greater role in the era of big data.