Application of Elastic Waves and Neural Networks for the Prediction of Forces in Bolts of Flange Connections Subjected to Static Tension Tests

There is a group of measurement techniques that can be used in the task of force identification in steel bolts. In this paper, the potential of elastic wave propagation signals was studied for possible application in force monitoring systems. A series of laboratory tests was carried out on flange connections subjected to static tensile tests. Each one contained six screws of the same diameter. Four bolts were equipped with washer load cells. Alternatively, selected bolts were equipped with piezoelectric transducers (actuator and sensor) in order to measure the elastic wave signals. Principal components analysis, time of arrival, and neural network compression were used for dimensionality reduction of the measured signals. Examples of the obtained results with respect to the studied connections show that the tension forces in bolts can be estimated with relatively good accuracy.