<title>Computer interpretation of ultrasonic pulse-echo signals for concrete dams</title>

Abstract

The nation's failing infrastructure points to the need to develop new techniques and equipment for the diagnostic evaluation of concrete in locks, dams, bridges, pavements, etc. An ultrasonic pulse echo (UPE) system for the non-destructive testing of concrete structures was developed at the U.S. Army Corps of Engineers, Waterways Experiment Station (WES) in the late 1980s. Because of the overwhelming amount of information in ultrasonic pulse echo (UPE) signals from concrete, learning to interpret raw data can be confusing. Although a human has the ability to recognize and calculate the complex signal information it can take months or years to master interpretation. The application of digital signal processing (DSP) algorithms for computerized signal interpretation can help reduce these problems. Computer processing of signals is also needed so that there is less dependency on highly experienced personnel and more objective diagnostic decisions can be made. This paper explores several DSP techniques aimed at computer interpretation. Three techniques were investigated: split- spectrum processing (SSP), ray-based modeling (RAM), and artificial neural networks (ANN).