On the use of the GP-NARX model for predicting hysteresis effects of bolted joint structures

Structures joined by lap-joints can present complex nonlinear dynamic behavior as a function of the stress to which lap-joint is subjected, including contact stiffness variations and softening, along with hysteresis effects related frictional dissipation at interface. Considering applications where use non-parametric models that depend only on input output data required, this work proposes details GP-NARX model’s approximate systems’ dynamics hysteresis. Initially, proposed predictive applicability evaluated numerical application involving Bouc-Wen oscillator hysteretic damping. Then, model describe BERT benchmark, an experimental system contains symmetric double bolted joint nonlinearly dependent upon applied excitation amplitudes, presenting friction joint’s well-known softening effect. The structure also presents variation presence uncertainties in measurement process. Thus, accommodate variability, training step considers several realizations. results indicate make accurate predictions response both investigated applications, emphasizing its practical ability, confidence intervals were able noisy data, learning relation between points.