An improved system for measuring tensions in multiple bolts is based on the usual measurement of the times of propagation of ultrasonic waves along the bolts, but the quantities measured and the method of processing the measurement data differ from those of conventional ultrasonic bolt-tension gauges. Typically, a conventional ultrasonic bolt-tension gauge utilizes a single particular feature of the ultrasonic waveform (e.g., a single zero-crossing) to measure the round-trip propagation time. When the gauge is functioning correctly, it can measure time with sufficient accuracy to give the bolt tension within ±2 percent. However, sometimes, the gauge focuses on the wrong signal feature (e.g., the wrong zero-crossing), yielding a reading that can be erroneous by as much as ±30 percent.
The system was initially designed for remote measurement of the tensions in several bolts in an article subjected to a potentially dangerous pressure test. The system includes a commercial ultrasonic bolt gauge with a microprocessor that serves as a gauge controller, plus serial data links between the microprocessor and host computer located in a safe control room remote from the test article. Under control by the computer, the microprocessor causes the ultrasonic bolt gauge to sequentially address ultrasonic transducers on individual bolts and transmit digitized responses to the computer.
Ideally, one would extract maximum information by utilizing the entire ultrasonic waveform rather than only a single feature of the waveform. The approach taken in designing the improved system was to reduce the extent and probability of error by use of a signal-analysis technique, utilizing the full-waveform approach. The improved system implements a correlation technique and also the times of multiple zero-crossings of the ultrasonic waveform for each bolt. The identification of zero-crossings and other signal features is enhanced by use of waveform-feature-recognition software based on three independent mathematical models of bolt gauges. The basic time-interval measurement is obtained from cross-correlating the tensioned and untensioned waveforms. A double- and triple-check is obtained from the zero crossings. If the intervals agree, then the time measured on the waveform for each bolt is considered to be more reliable than if the intervals do not agree. If the intervals do not agree, a vote can be taken and the waveform that does not agree can be discarded. Alternatively, the results can be averaged to obtain a final result that is less erroneous in the worst case.
This work was done by Stuart M. Gleman of I-NET and Lyle J. Robinson, Stephen W. Thayer, Geoffrey K. Rowe, David L. Thompson, and Carl G. Hallberg of Dynacs Engineering Co., Inc., for Kennedy Space Center. For further information, access the Technical Support Package (TSP) free on-line at www.techbriefs.com under the Mechanics category, or circle no. 192 on the TSP Order Card in this issue to receive a copy by mail ($5 charge).
This invention is owned by NASA, and a patent application has been filed. Inquiries concerning nonexclusive or exclusive license for its commercial development should be addressed to the Patent Counsel, Kennedy Space Center; (407) 867-6225.