Modeling and characterization of flexible polyurethane foam

Rong Deng, Purdue University

Abstract

Hereditary and fractional derivative models are used to study the dynamic and quasi-static properties of flexible polyurethane foams used in automotive seat cushions. Nonlinear elastic and linear viscoelastic properties are incorporated into these two models. A polynomial function of the compression is used to represent the nonlinear elastic behavior. The viscoelastic property is modeled by a convolution of the displacement and a relaxation kernel consisting of exponential terms in the hereditary model, and by differential equations involving fractional derivative terms in the fractional derivative model. The analytical relationship between the two models is discussed. The models are used to predict the responses of s system composed of a mass supported by polyurethane foam, identified as a foam-block system, that undergoes uni-axial displacements in two situations: impulse excitation and harmonic excitation around a pre-compression point. The models are also applied to explain the response characteristics in quasi-static compression tests of the foam blocks. In each of these cases system identification procedures are formulated and effects of imperfections, such as dry friction and measurement noise, are examined through extensive sets of simulations. While good agreement between the responses measured in experiments and the responses predicted from the estimated models was often obtained, some of the model parameters changed significantly when the excitation characteristics or mean foam compression changed. This indicates the need for more complex models. In general the more flexibility offered by the hereditary model resulted in better performance. For the quasi-static tests over multiple cycles, multi-component fractional derivative models were required to predict the multi-cycle response behavior. Recommendations for future work include investigation of system identification and experimental techniques to estimate the parameters of multi-component viscoelastic models, which are clearly necessary if we wish to extend the range of applicability of individual models.

Degree

Ph.D.

Advisors

Bajaj, Purdue University.

Subject Area

Mechanical engineering

COinS