The applied of self-organizing clustering analysis on Coin-tap Test system of airplane composite structure

Coin-tap Test is a kind of NDT methods used commonly. The test is restricted by composite material structure less, but needs a large amount of data. To solve the difficulty of making test pieces and complicated issues of Coin-tap Test data processing, we put forward the clustering analysis of self-organizing neural network to deal with Coin-tap data. With the aid of MATLAB toolbox, the method applied to Coin-tap Test succeeded in hitting data are classified and finding out the damage location. Compared the clustering analysis results with the actual damage, the accuracy of data can reach 90% in 32 groups, and the accuracy is higher with the greater amount of data. Therefore, clustering analysis of self-organizing neural network has a good stability and high precision. Combined with Coin-tap test method, this analysis can basically meet the requirements of aircraft composite nondestructive testing, and it has a good prospect of engineering application.