Genetic algorithms and neural networks in optimal location of piezoelectric actuators and identification of mechanical properties

  • Authors:

  • L. Roseiro;U. Ramos;R. Leal

  • Affiliations:

  • Instituto Superior de Engenharia de Coimbra, Coimbra, Portugal;Instituto Superior de Engenharia de Coimbra, Coimbra, Portugal;DEM, CEMUC, FCT, Universidade de Coimbra, Coimbra, Portugal

  • Venue:
  • SMO'06 Proceedings of the 6th WSEAS International Conference on Simulation, Modelling and Optimization
  • Year:
  • 2006

Quantified Score

Hi-index 0.00

Visualization

Abstract

Composite materials are a very important class of engineering materials with great properties and applications in a variety of complex structures. The correct design of such structures requires adequate analysis and, in particular, adequate and accurate models for simulation with numerical tools. Piezoelectric materials can convert mechanical energy into electric energy and vice-versa and are being applied in laminated composite structures, working either as sensors (mechanical load applications) or actuators (electric potential applications), allowing its use in a wide range of engineering applications. Two important subject matters in the type of applications involving laminated composite structures and piezoelectric materials are the identification of material constants and the optimal location of actuators and sensors. In this work two numerical procedures involving genetic algorithms and neural networks are proposed for these problems. A neural networks based methodology for the identification of mechanical properties is presented. The identification process makes use of the information collected from piezoelectric sensors. Pairs of sensors are placed on the surfaces of a composite laminated plate and the potential differences are obtained among pairs when the plate is charged. Since genetic algorithms are very expensive when the objective function has high computational cost, we introduce the artificial neural networks to improve the efficiency of the genetic algorithm for the optimal location of piezoelectric actuators. Neural networks make a choice of the chromosomes for which it is worthwhile to calculate the objective function; for the other chromosomes neural networks attribute a value to the objective function. For both procedures, the results obtained are compared with those present in literature.