LS-SVM Regression for Structural Damage Diagnosis Using the Iterated Improved Reduction System

Abstract

A damage detection and estimation method is proposed for structural health monitoring using incomplete modal data and least squares support vector machine (LS-SVM). To accommodate the use of incomplete modal data, the iterated improved reduction system (IIRS) method has been used to condense the mass and sti®ness matrices of the structure. The¯rst two incomplete mode shapes and natural frequencies of a damaged structure are used as input data to the LS-SVM. The coupled simulated annealing (CSA) and standard simplex method using 10-fold cross-validation techniques are adopted to determine the optimal tuning parameters in the LS-SVM model. Three illustrative examples with and without noise in modal data are prepared to evaluate the performance of the proposed method. The results indicated that this method can be reliably used to identify the damages of structures with good accuracy. Int. J. Str. Stab. Dyn. Downloaded from www.worldscientific.com by FUDAN UNIVERSITY on 05/15/15. For personal use only.

Research on the relationship between the scattering contribution and physical factors of the reference radiation regulated by ISO 4037–1

“…However, the solution of quadratic programming is complicated, the method is usually combined with the least square method. The inequality constraints are transformed into equality constraints, and the complex nonlinear solution process is transformed into a linear matrix solution under the premise of ensuring accuracy, which expands the usability of the model [ 15 ].…”

Review on the new development of vibration-based damage identification for civil engineering structures: 2010–2019

“…Kourehli [175] used the first two incomplete mode shapes and natural frequencies as input data to train the SVM. In this approach, a radical basis function (RBF) was chosen as a kernel function.…”

Review on Vibration-Based Structural Health Monitoring Techniques and Technical Codes

“…Recently, machine learning methods become more popular due to the quick development of artificial intelligence [50][51][52][53][54][55][56][57][58]. It can definitely help to improve the reconstruction of structural model, but the model is a data-driven model rather than the physics-based model in FE model updating methods.…”