Development of crack detection and length estimation software system.

摘要

Structural integrity is an important topic in the discussion of aircrafts. As an aircraft is in flight it undergoes a variety of structural disturbances. The occurrences can create undue stress on the structure of the aircraft which can impact flight safety, stability, and even the efficiency of the engines. Structural Health Monitoring is the process of implementing a damage detection and characterization strategy for engineering structures. In this thesis, a crack detection and crack length estimation software system was developed, implemented and tested. The feature extraction method for the crack detection was chosen by the University of Dayton Research Institute (UDRI) and it fit the proposed technique of statistical analysis for feature extraction for this thesis. The feature extraction for crack length estimation was developed and tested in this thesis. The crack detection subsystem was developed using extracted features provided by the UDRI and were implemented using an artificial neural network on MATLAB software environment. The crack length estimation subsystem was developed using correlation coefficient for feature extraction and were implemented using adaptive neuro-fuzzy inference system on MATLAB software environment. The experimental data was then used to test the accuracy of the crack detection and crack length estimation software system. The developed software system was tested for accuracy on four aluminum dog-bone coupons to simulate the structure of an aircraft. The data used to simulate a crack was provided by the UDRI in conjunction with the AFRL. The crack detection subsystem testing produced an accuracy of detection of 85% or better for all coupons tested. The crack length estimation subsystem testing provided an 85% accuracy of estimation with an error no greater than 5% of estimation. The data was tested for error then divided by the number of actual recorded lengths. The observed results from testing show that the requirements were met as expected. The system developments, the simulation results and testing of the crack detection and crack length estimation software will be presented in this thesis report.