Noise Analysis And Studying Of Adaptive Noise Cancellation In Pipeline Detecting System

In the processing of producing, various industrial pipelines must be test regularly. It is an urgently issue to be solved that determining the defects expansion precisely and then predicting expectancy accurately. In recent years, MFL Detecting equipments are widely used in pipeline quality inspection. It can implement visualization, without blind spots and automated dynamic testing. However, it usually can only detect defects' location rather than type and the accuracy rate is not high. The basic reason is that those methods can't effectively extract the useful signal from the noise. It can't meet the requirements of identifying defects' type.Currently, traditional methods of canceling the noise are low-pass filter, software and the improvement of mechanical structure and so on. These methods can reduce noise and interference to some extent, but the result is not satisfactory to meet the requirements of the follow-up signal processing.In this paper, firstly, we analysis noise sources in the detecting system. According it we proposed some effective methods. For example, for system noise, low-noise circuit should be designed. For outside interference three categories of interference suppression methods are present.Then we discussed the theory of adaptive filter. The performance of LMS, NLMS, and RLS algorithm have been analyzed in detail. By comparing, we draw some useful conclusions. Finally adaptive noise cancellation systems is designed to processing MFL signal. The results proved this system can effectively eliminate 50Hz noise and interference when the two road of noise signals are line relevant.In order to reduce the complexity of detection system, we designed software filter-adaptive notch filter. The test shows that it can effectively track the frequency changes of interference to eliminate the impact of frequency interference.Due to the limitation of the traditional experimental methods, which need too many instruments, too long time, and have difficulties for system design and adjustment, a new experimental method based on Simulink toolbox of Matlab has been put forward. With it the adaptive noise canceller based on Artificial Neural Network have been designed. The results proved it's efficient in the noise cancellation.