Research of Acoustic Early Warning Detection Technology for the Pipeline Leak Based on Blind Source Separation
|Course||Power Electronics and Power Drives|
|Keywords||Leak warning Blind Source Separation Maximum Entropy Signal analysis|
Pipeline safety detection technology is a key technology that is closely related with the national economy. Currently, most of our pipeline spills are caused by a punch of Stolen and mechanical excavation. Therefore, how to prevent man-made destruction the pipeline early warning detection technology is an urgent need to resolve the core technology of our pipeline transportation industry, the technology of production safety of the pipeline has a very important practical significance. This paper proposes a blind source separation technology pipeline leakage acoustic warning detection methods, the use of early warning detection of pipeline eigenvalue damage signal detection can be realized through the pipeline mixing acoustic signal unmixing, real restore corrupt the signal, . Maximum Entropy algorithm based on the natural gradient, designed and implemented a real-time two-dimensional hybrid acoustic blind source separation system successfully achieve the separation of the two routes within the pipe model instantaneous mixture of acoustic signals. Propose and confirm a new pipeline leak warning technology research. In this paper, do the following aspects work: (1) the blind source separation theory is applied to the pipeline acoustic leak detection technology issues of key research pipeline, two-channel linear instantaneous mixture separation of acoustic signals. Proposed a two line the instantaneous mixing acoustic blind source separation algorithm based on the natural gradient of the maximum entropy method. Square wave and triangular wave mixing - separating simulation confirmed the effectiveness of the algorithm, the algorithm has a certain robustness, and low noise. (2) The design and implementation of a real-time two-dimensional hybrid acoustic blind source separation system. The sampling frequency of the system, the algorithm is simple, the two routes instantaneous mixing acoustic signal acquisition, separation and signal analysis and data storage. Learning rate optimization to further improve the system's real-time signal acquisition and data processing and analysis capabilities. Provides for the application of the theory practice test and analysis platform. (3) the use of real-time two-dimensional hybrid acoustic separation system to carry out the blind source separation signal aliasing loop simulation. Through a pilot study successfully achieved the destruction of the pipe model with vibration linear instantaneous mixture separation of acoustic signals. The time-frequency domain analysis showed that the isolated destruction signal having a visible identification characteristics in the pipeline early warning detection has a high application value. Verified based on blind source separation of the feasibility and effectiveness of the pipeline leak sound detection technology In summary, this paper explores a new pipeline leak detection method, and confirmed the correctness and feasibility of the method used to test the data analysis provide a solid theoretical basis and practical experience for the follow-up in-depth study of the piping sound detection technology based on blind source separation.