The Research Of Acoustic Emission Source Location Method To Tank Bottom Corrosion

Because corrosion can lead to many kinds of phenomenon of pitting corrosion, oxidation film falling and crack on the tank bottom sheet, if not found in time and properly disposed, corrosion will bring huge security risk to the tank. Compared to traditional tank bottom detection methods, acoustic emission detection technology can give on-line monitoring without opening tank and significant savings in detection costs. In the tank bottom corrosion acoustic emission detection, acoustic emission source location is one of the main parts, providing important basis for maintenance at the tank bottom.This paper focuses on the research of acoustic emission source location technology in acoustic emission source clustering, time difference measurement, positioning algorithm. In the acoustic emission signal detection of the tank bottom, the sensor will receive acoustic emission signals from different sources, in order to ensure accurate positioning of the source location, we must distinguish between acoustic emission signals from different sources. High degree of similarity between signals from the homologous acoustic emission can be used as an important basis to distinguish between different source signals. The fuzzy C-means cluster is a kind of clustering algorithm that has rapid convergence and ability of handling large data sets, commonly used in the feature analysis, classifier design, data identification, but classification number must be known in advance. This paper uses fuzzy subtractive cluster to determine the number of categories to offset the lack of priori information in the use of fuzzy C-means cluster, and improve the application of fuzzy C-means cluster.Compared to zone location, TDOA location has high location accuracy, easier to use than the energy location, and its calculation is simple, this paper adopts the least time scale variance method to achieve TDOA location. The least time scale variance method gets rid of the consideration to the geometric relationships as triangulation method, converts the calculation of the acoustic emission source coordinates into solving constrained nonlinear optimization problem, and makes the algorithm more simple and accurate. In view of the nonlinear and non-stationary characteristics in acoustic emission signals, spectrum entropy energy product was introduced into the acoustic emission signal time difference measurement, and the effectiveness of the proposed method was verified by experiments. In the verification experiments, the method of increasing the number of sensors was used to erase velocity parameter in location formulas, so that the location results only related to the time difference measurement.Designed the simulation experiment of tank bottom corrosion acoustic emission location; determined the acoustic emission signal sources by fuzzy C-means cluster; used spectrum entropy energy product algorithm to measure the time difference; obtained location results within the scope of permissible error by the use of the least time scale variance method; finally analyzed the cause of the error combining with location data and error comparison.