I have been attempting to simulate the sliding of masses at floor levels for a four-story structure using SAP 2000's 'Plastic Wen' link element. The hysteresis loops showed a reasonably accurate pattern. The acceleration at various shelf levels, on the other hand, exhibits incredibly sharp peaks for a very short duration (high-frequency noise). To eliminate the noise, I have increased the damping ratio for higher modes, increased the HHT alpha value and defined mass and rotational inertia for link elements. Furthermore, the elastic stiffness of the Bouc-Wen element was reduced to a lower value which still preserves the rigid plastic shape of the hysteresis loops. Nothing pretty much eliminated the acceleration response's high-frequency noise. Could someone shed some light on how to resolve these?