Optimization of structural health monitoring attributes under variable failure rate condition using teaching learning based optimization and multiple regression

Abstract

Abstract The manufacturing industries prefer to use exponential equations for evaluating reliability, failure rate, etc., while assuming failure rate constant, but in real life failure rate is not always constant. This research work proposes a novel combination of Multiple Regression and Teaching Learning Based Optimization (TLBO). This work starts with regression modeling, to form an objective function from available data; secondly TLBO technique used, to obtain optimum values of structural health monitoring (SHM) attributes through cost minimization and finally, Weibull analysis to check the effect of optimum values of SHM attributes on β. The final results indicated that the β parameter reduces, which symbolized reduction in failure rate & SHM cost and, improvement in system reliability, thus validates the adopted methodology of combing multiple regression with TLBO, thus success of this work become the backbone for intelligent structures for SHMsystem.