Electro-structural analysis and optimization studies of laminated composite beam energy harvester

Abstract

Modern ambient energy harvesters meet the micropower demands of the industrial sensors and communication hardware. Optimum design of such systems to maximize the power output is, therefore, of vital importance. This paper presents a multi-objective optimization methodology of smart laminated composite beam energy harvester by considering different geometric parameters. A typical multilayer carbon-epoxy composite beam with unimorph piezoelectric configuration is analyzed with finite element model. The coupled electro-mechanical effects are considered and the resultant free vibration solution is validated with that of a three-dimensional model. The structural response and power density of the base excited beam are obtained by varying thickness ratio of piezo layer, composite lamination sequence, number of layers, circuit resistance, piezoelectric materials, and so on. The most influencing design variables are identified using analysis of variance (ANOVA). By minimization of the frequency band gap between first two modes and maximization of output power density simultaneously, optimum solution is obtained from grey relation analysis technique. The results are validated with a surrogate model employing multilayer perceptron (MLP) neural network in conjunction with multi-objective genetic algorithms. The structural dynamic response and electrical power output characteristics of optimized configuration are found to be improved.

Keywords:

• Artificial neural networks
• energy harvesting
• finite element modeling
• genetic algorithms
• Grey relational analysis
• laminated composite beam
• piezoelectric patch