Date of Graduation


Document Type


Degree Type



Statler College of Engineering and Mineral Resources


Civil and Environmental Engineering

Committee Chair

Roger H. L. Chen.


Fiber Reinforced Polymer composites are relatively new to the bridge construction industry. One of the challenges to greater usage is the lack of a real-time nondestructive method to accurately evaluate them. A Neural Network program can be used to predict the status of a structure by analyzing summarized Acoustic Emission data from the structure in real-time.;Existing methods use tools such as MS Excel to extract the Neural Network input matrix from raw data having more than 30,000 rows of Acoustic Emission data for a simple three point bending test. This manual data extraction process is very time consuming. Therefore, application software was developed to efficiently summarize the Acoustic Emission data into a Neural Network input matrix. The awk and C scripts were used to develop the application. The awk programming language is designed to search for, match patterns, and perform actions on text files. The awk programs are generally quite small, easy to understand and are easily interpreted. This makes it a good pattern matching and data retrieval language. An awk program can be executed by using a gcc compiler, so a C module was developed to combine all the awk scripts into a single application. The application software was used to extract the Neural Network input matrix from previously conducted AE experiments; a comparison of manually prepared matrices and those prepared by application is presented.;Tension, Bending and Fatigue experiments of FRP specimens were conducted and the AE data obtained from the experiments were used to analyze the structural properties of the specimens with the help of the application software developed through this study. Loading quarter of the specimens was predicted using the Neural Network program. The predictions obtained from Neural Network program for the matrices prepared by the application software were found to be more accurate and consistent than the predictions from the manually prepared matrices.