Modeling MR-dampers: The ridgenet estimation approach

This paper develops a new technique, the ridgenet estimation, for the blackbox modeling of MR-dampers. It continues the original effort on wavelets-based identification (Jin et al [1]) and makes a step forward by employing the ridgelet network to construct the nonlinear mapping. The paper starts with an introduction section where we present an overview on the previous MR-damper modeling, the recent result on ridgelet analysis, and the outline of the proposed approach. Then the major content is divided into two sections. In Section 2, we give a detailed description of the ridgenet algorithm in the form a three-step computational procedure: the ridgelet dictionary construction, the ridgelet basis selection, and the ridgelet network optimization. In Section 3, the numerical implementation of the ridgenet algorithm is addressed and the application result is reported. We conclude in Section 4 and point out a direction to expand the current result.